fix(deep-crawl): BestFirst priority inversion; remove pre-scoring truncation. ref #1253

Use negative scores in PQ to visit high-score URLs first and drop link cap prior to scoring; add test for ordering.

.github/workflows/main.yml

@@ -9,16 +9,26 @@ on:
     types: [opened]
   discussion:
     types: [created]
+  watch:
+    types: [started]
 
 jobs:
   notify-discord:
     runs-on: ubuntu-latest
     steps:
+      - name: Send to Google Apps Script (Stars only)
+        if: github.event_name == 'watch'
+        run: |
+          curl -fSs -X POST "${{ secrets.GOOGLE_SCRIPT_ENDPOINT }}" \
+            -H 'Content-Type: application/json' \
+            -d '{"url":"${{ github.event.sender.html_url }}"}'
       - name: Set webhook based on event type
         id: set-webhook
         run: |
           if [ "${{ github.event_name }}" == "discussion" ]; then
             echo "webhook=${{ secrets.DISCORD_DISCUSSIONS_WEBHOOK }}" >> $GITHUB_OUTPUT
+          elif [ "${{ github.event_name }}" == "watch" ]; then
+            echo "webhook=${{ secrets.DISCORD_STAR_GAZERS }}" >> $GITHUB_OUTPUT
           else
             echo "webhook=${{ secrets.DISCORD_WEBHOOK }}" >> $GITHUB_OUTPUT
           fi
@@ -31,5 +41,6 @@ jobs:
           args: |
             ${{ github.event_name == 'issues' && format('📣 New issue created: **{0}** by {1} - {2}', github.event.issue.title, github.event.issue.user.login, github.event.issue.html_url) || 
             github.event_name == 'issue_comment' && format('💬 New comment on issue **{0}** by {1} - {2}', github.event.issue.title, github.event.comment.user.login, github.event.comment.html_url) ||
-            github.event_name == 'pull_request' && format('🔄 New PR opened: **{0}** by {1} - {2}', github.event.pull_request.title, github.event.pull_request.user.login, github.event.pull_request.html_url) || 
+            github.event_name == 'pull_request' && format('🔄 New PR opened: **{0}** by {1} - {2}', github.event.pull_request.title, github.event.pull_request.user.login, github.event.pull_request.html_url) ||
+            github.event_name == 'watch' && format('⭐ {0} starred Crawl4AI 🥳! Check out their profile: {1}', github.event.sender.login, github.event.sender.html_url) ||
             format('💬 New discussion started: **{0}** by {1} - {2}', github.event.discussion.title, github.event.discussion.user.login, github.event.discussion.html_url) }}

.github/workflows/release.yml

@@ -0,0 +1,142 @@
+name: Release Pipeline
+on:
+  push:
+    tags:
+      - 'v*'
+      - '!test-v*'  # Exclude test tags
+
+jobs:
+  release:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write  # Required for creating releases
+    
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+      
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+      
+      - name: Extract version from tag
+        id: get_version
+        run: |
+          TAG_VERSION=${GITHUB_REF#refs/tags/v}
+          echo "VERSION=$TAG_VERSION" >> $GITHUB_OUTPUT
+          echo "Releasing version: $TAG_VERSION"
+      
+      - name: Install package dependencies
+        run: |
+          pip install -e .
+      
+      - name: Check version consistency
+        run: |
+          TAG_VERSION=${{ steps.get_version.outputs.VERSION }}
+          PACKAGE_VERSION=$(python -c "from crawl4ai.__version__ import __version__; print(__version__)")
+          
+          echo "Tag version: $TAG_VERSION"
+          echo "Package version: $PACKAGE_VERSION"
+          
+          if [ "$TAG_VERSION" != "$PACKAGE_VERSION" ]; then
+            echo "❌ Version mismatch! Tag: $TAG_VERSION, Package: $PACKAGE_VERSION"
+            echo "Please update crawl4ai/__version__.py to match the tag version"
+            exit 1
+          fi
+          echo "✅ Version check passed: $TAG_VERSION"
+      
+      - name: Install build dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install build twine
+      
+      - name: Build package
+        run: python -m build
+      
+      - name: Check package
+        run: twine check dist/*
+      
+      - name: Upload to PyPI
+        env:
+          TWINE_USERNAME: __token__
+          TWINE_PASSWORD: ${{ secrets.PYPI_TOKEN }}
+        run: |
+          echo "📦 Uploading to PyPI..."
+          twine upload dist/*
+          echo "✅ Package uploaded to https://pypi.org/project/crawl4ai/"
+      
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+      
+      - name: Log in to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_TOKEN }}
+      
+      - name: Extract major and minor versions
+        id: versions
+        run: |
+          VERSION=${{ steps.get_version.outputs.VERSION }}
+          MAJOR=$(echo $VERSION | cut -d. -f1)
+          MINOR=$(echo $VERSION | cut -d. -f1-2)
+          echo "MAJOR=$MAJOR" >> $GITHUB_OUTPUT
+          echo "MINOR=$MINOR" >> $GITHUB_OUTPUT
+      
+      - name: Build and push Docker images
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          push: true
+          tags: |
+            unclecode/crawl4ai:${{ steps.get_version.outputs.VERSION }}
+            unclecode/crawl4ai:${{ steps.versions.outputs.MINOR }}
+            unclecode/crawl4ai:${{ steps.versions.outputs.MAJOR }}
+            unclecode/crawl4ai:latest
+          platforms: linux/amd64,linux/arm64
+      
+      - name: Create GitHub Release
+        uses: softprops/action-gh-release@v2
+        with:
+          tag_name: v${{ steps.get_version.outputs.VERSION }}
+          name: Release v${{ steps.get_version.outputs.VERSION }}
+          body: |
+            ## 🎉 Crawl4AI v${{ steps.get_version.outputs.VERSION }} Released!
+            
+            ### 📦 Installation
+            
+            **PyPI:**
+            ```bash
+            pip install crawl4ai==${{ steps.get_version.outputs.VERSION }}
+            ```
+            
+            **Docker:**
+            ```bash
+            docker pull unclecode/crawl4ai:${{ steps.get_version.outputs.VERSION }}
+            docker pull unclecode/crawl4ai:latest
+            ```
+            
+            ### 📝 What's Changed
+            See [CHANGELOG.md](https://github.com/${{ github.repository }}/blob/main/CHANGELOG.md) for details.
+          draft: false
+          prerelease: false
+          token: ${{ secrets.GITHUB_TOKEN }}
+      
+      - name: Summary
+        run: |
+          echo "## 🚀 Release Complete!" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 📦 PyPI Package" >> $GITHUB_STEP_SUMMARY
+          echo "- Version: ${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY
+          echo "- URL: https://pypi.org/project/crawl4ai/" >> $GITHUB_STEP_SUMMARY
+          echo "- Install: \`pip install crawl4ai==${{ steps.get_version.outputs.VERSION }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 🐳 Docker Images" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:${{ steps.get_version.outputs.VERSION }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:${{ steps.versions.outputs.MINOR }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:${{ steps.versions.outputs.MAJOR }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:latest\`" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 📋 GitHub Release" >> $GITHUB_STEP_SUMMARY
+          echo "https://github.com/${{ github.repository }}/releases/tag/v${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY

.github/workflows/test-release.yml.disabled

@@ -0,0 +1,116 @@
+name: Test Release Pipeline
+on:
+  push:
+    tags:
+      - 'test-v*'
+
+jobs:
+  test-release:
+    runs-on: ubuntu-latest
+    
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+      
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+      
+      - name: Extract version from tag
+        id: get_version
+        run: |
+          TAG_VERSION=${GITHUB_REF#refs/tags/test-v}
+          echo "VERSION=$TAG_VERSION" >> $GITHUB_OUTPUT
+          echo "Testing with version: $TAG_VERSION"
+      
+      - name: Install package dependencies
+        run: |
+          pip install -e .
+      
+      - name: Check version consistency
+        run: |
+          TAG_VERSION=${{ steps.get_version.outputs.VERSION }}
+          PACKAGE_VERSION=$(python -c "from crawl4ai.__version__ import __version__; print(__version__)")
+          
+          echo "Tag version: $TAG_VERSION"
+          echo "Package version: $PACKAGE_VERSION"
+          
+          if [ "$TAG_VERSION" != "$PACKAGE_VERSION" ]; then
+            echo "❌ Version mismatch! Tag: $TAG_VERSION, Package: $PACKAGE_VERSION"
+            echo "Please update crawl4ai/__version__.py to match the tag version"
+            exit 1
+          fi
+          echo "✅ Version check passed: $TAG_VERSION"
+      
+      - name: Install build dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install build twine
+      
+      - name: Build package
+        run: python -m build
+      
+      - name: Check package
+        run: twine check dist/*
+      
+      - name: Upload to Test PyPI
+        env:
+          TWINE_USERNAME: __token__
+          TWINE_PASSWORD: ${{ secrets.TEST_PYPI_TOKEN }}
+        run: |
+          echo "📦 Uploading to Test PyPI..."
+          twine upload --repository testpypi dist/* || {
+            if [ $? -eq 1 ]; then
+              echo "⚠️ Upload failed - likely version already exists on Test PyPI"
+              echo "Continuing anyway for test purposes..."
+            else
+              exit 1
+            fi
+          }
+          echo "✅ Test PyPI step complete"
+      
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+      
+      - name: Log in to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_TOKEN }}
+      
+      - name: Build and push Docker test images
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          push: true
+          tags: |
+            unclecode/crawl4ai:test-${{ steps.get_version.outputs.VERSION }}
+            unclecode/crawl4ai:test-latest
+          platforms: linux/amd64,linux/arm64
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+      
+      - name: Summary
+        run: |
+          echo "## 🎉 Test Release Complete!" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 📦 Test PyPI Package" >> $GITHUB_STEP_SUMMARY
+          echo "- Version: ${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY
+          echo "- URL: https://test.pypi.org/project/crawl4ai/" >> $GITHUB_STEP_SUMMARY
+          echo "- Install: \`pip install -i https://test.pypi.org/simple/ crawl4ai==${{ steps.get_version.outputs.VERSION }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 🐳 Docker Test Images" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:test-${{ steps.get_version.outputs.VERSION }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "- \`unclecode/crawl4ai:test-latest\`" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "### 🧹 Cleanup Commands" >> $GITHUB_STEP_SUMMARY
+          echo "\`\`\`bash" >> $GITHUB_STEP_SUMMARY
+          echo "# Remove test tag" >> $GITHUB_STEP_SUMMARY
+          echo "git tag -d test-v${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY
+          echo "git push origin :test-v${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY
+          echo "" >> $GITHUB_STEP_SUMMARY
+          echo "# Remove Docker test images" >> $GITHUB_STEP_SUMMARY
+          echo "docker rmi unclecode/crawl4ai:test-${{ steps.get_version.outputs.VERSION }}" >> $GITHUB_STEP_SUMMARY
+          echo "docker rmi unclecode/crawl4ai:test-latest" >> $GITHUB_STEP_SUMMARY
+          echo "\`\`\`" >> $GITHUB_STEP_SUMMARY

.gitignore

@@ -1,3 +1,6 @@
+# Scripts folder (private tools)
+.scripts/
+
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]

CHANGELOG.md

@@ -21,6 +21,21 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+- **Flexible LLM Provider Configuration** (Docker): 
+  - Support for `LLM_PROVIDER` environment variable to override default provider
+  - Per-request provider override via optional `provider` parameter in API endpoints
+  - Automatic provider validation with clear error messages
+  - Updated Docker documentation and examples
+
+### Changed
+- **WebScrapingStrategy Refactoring**: Simplified content scraping architecture
+  - `WebScrapingStrategy` is now an alias for `LXMLWebScrapingStrategy` for backward compatibility
+  - Removed redundant BeautifulSoup-based implementation (~1000 lines of code)
+  - `LXMLWebScrapingStrategy` now inherits directly from `ContentScrapingStrategy`
+  - All existing code using `WebScrapingStrategy` continues to work without modification
+  - Default scraping strategy remains `LXMLWebScrapingStrategy` for optimal performance
+
 ### Added
 - **AsyncUrlSeeder**: High-performance URL discovery system for intelligent crawling at scale
   - Discover URLs from sitemaps and Common Crawl index

Dockerfile

@@ -1,7 +1,7 @@
 FROM python:3.12-slim-bookworm AS build
 
 # C4ai version
-ARG C4AI_VER=0.6.0
+ARG C4AI_VER=0.7.0-r1
 ENV C4AI_VERSION=$C4AI_VER
 LABEL c4ai.version=$C4AI_VER
 

PROGRESSIVE_CRAWLING.md

@@ -0,0 +1,320 @@
+# Progressive Web Crawling with Adaptive Information Foraging
+
+## Abstract
+
+This paper presents a novel approach to web crawling that adaptively determines when sufficient information has been gathered to answer a given query. Unlike traditional exhaustive crawling methods, our Progressive Information Sufficiency (PIS) framework uses statistical measures to balance information completeness against crawling efficiency. We introduce a multi-strategy architecture supporting pure statistical, embedding-enhanced, and LLM-assisted approaches, with theoretical guarantees on convergence and practical evaluation methods using synthetic datasets.
+
+## 1. Introduction
+
+Traditional web crawling approaches follow predetermined patterns (breadth-first, depth-first) without consideration for information sufficiency. This work addresses the fundamental question: *"When do we have enough information to answer a query and similar queries in its domain?"*
+
+We formalize this as an optimal stopping problem in information foraging, introducing metrics for coverage, consistency, and saturation that enable crawlers to make intelligent decisions about when to stop crawling and which links to follow.
+
+## 2. Problem Formulation
+
+### 2.1 Definitions
+
+Let:
+- **K** = {d₁, d₂, ..., dₙ} be the current knowledge base (crawled documents)
+- **Q** be the user query
+- **L** = {l₁, l₂, ..., lₘ} be available links with preview metadata
+- **θ** be the confidence threshold for information sufficiency
+
+### 2.2 Objectives
+
+1. **Minimize** |K| (number of crawled pages)
+2. **Maximize** P(answers(Q) | K) (probability of answering Q given K)
+3. **Ensure** coverage of Q's domain (similar queries)
+
+## 3. Mathematical Framework
+
+### 3.1 Information Sufficiency Metric
+
+We define Information Sufficiency as:
+
+```
+IS(K, Q) = min(Coverage(K, Q), Consistency(K, Q), 1 - Redundancy(K)) × DomainCoverage(K, Q)
+```
+
+### 3.2 Coverage Score
+
+Coverage measures how well current knowledge covers query terms and related concepts:
+
+```
+Coverage(K, Q) = Σ(t ∈ Q) log(df(t, K) + 1) × idf(t) / |Q|
+```
+
+Where:
+- df(t, K) = document frequency of term t in knowledge base K
+- idf(t) = inverse document frequency weight
+
+### 3.3 Consistency Score
+
+Consistency measures information coherence across documents:
+
+```
+Consistency(K, Q) = 1 - Var(answers from random subsets of K)
+```
+
+This captures the principle that sufficient knowledge should provide stable answers regardless of document subset.
+
+### 3.4 Saturation Score
+
+Saturation detects diminishing returns:
+
+```
+Saturation(K) = 1 - (ΔInfo(Kₙ) / ΔInfo(K₁))
+```
+
+Where ΔInfo represents marginal information gain from the nth crawl.
+
+### 3.5 Link Value Prediction
+
+Expected information gain from uncrawled links:
+
+```
+ExpectedGain(l) = Relevance(l, Q) × Novelty(l, K) × Authority(l)
+```
+
+Components:
+- **Relevance**: BM25(preview_text, Q)
+- **Novelty**: 1 - max_similarity(preview, K)
+- **Authority**: f(url_structure, domain_metrics)
+
+## 4. Algorithmic Approach
+
+### 4.1 Progressive Crawling Algorithm
+
+```
+Algorithm: ProgressiveCrawl(start_url, query, θ)
+  K ← ∅
+  crawled ← {start_url}
+  pending ← extract_links(crawl(start_url))
+  
+  while IS(K, Q) < θ and |crawled| < max_pages:
+    candidates ← rank_by_expected_gain(pending, Q, K)
+    if max(ExpectedGain(candidates)) < min_gain:
+      break  // Diminishing returns
+    
+    to_crawl ← top_k(candidates)
+    new_docs ← parallel_crawl(to_crawl)
+    K ← K ∪ new_docs
+    crawled ← crawled ∪ to_crawl
+    pending ← extract_new_links(new_docs) - crawled
+  
+  return K
+```
+
+### 4.2 Stopping Criteria
+
+Crawling terminates when:
+1. IS(K, Q) ≥ θ (sufficient information)
+2. d(IS)/d(crawls) < ε (plateau reached)
+3. |crawled| ≥ max_pages (resource limit)
+4. max(ExpectedGain) < min_gain (no promising links)
+
+## 5. Multi-Strategy Architecture
+
+### 5.1 Strategy Pattern Design
+
+```
+AbstractStrategy
+  ├── StatisticalStrategy (no LLM, no embeddings)
+  ├── EmbeddingStrategy (with semantic similarity)
+  └── LLMStrategy (with language model assistance)
+```
+
+### 5.2 Statistical Strategy
+
+Pure statistical approach using:
+- BM25 for relevance scoring
+- Term frequency analysis for coverage
+- Graph structure for authority
+- No external models required
+
+**Advantages**: Fast, no API costs, works offline
+**Best for**: Technical documentation, specific terminology
+
+### 5.3 Embedding Strategy (Implemented)
+
+Semantic understanding through embeddings:
+- Query expansion into semantic variations
+- Coverage mapping in embedding space
+- Gap-driven link selection
+- Validation-based stopping criteria
+
+**Mathematical Framework**:
+```
+Coverage(K, Q) = mean(max_similarity(q, K) for q in Q_expanded)
+Gap(q) = 1 - max_similarity(q, K)
+LinkScore(l) = Σ(Gap(q) × relevance(l, q)) × (1 - redundancy(l, K))
+```
+
+**Key Parameters**:
+- `embedding_k_exp`: Exponential decay factor for distance-to-score mapping
+- `embedding_coverage_radius`: Distance threshold for query coverage
+- `embedding_min_confidence_threshold`: Minimum relevance threshold
+
+**Advantages**: Semantic understanding, handles ambiguity, detects irrelevance
+**Best for**: Research queries, conceptual topics, diverse content
+
+### 5.4 Progressive Enhancement Path
+
+1. **Level 0**: Statistical only (implemented)
+2. **Level 1**: + Embeddings for semantic similarity (implemented)
+3. **Level 2**: + LLM for query understanding (future)
+
+## 6. Evaluation Methodology
+
+### 6.1 Synthetic Dataset Generation
+
+Using LLM to create evaluation data:
+
+```python
+def generate_synthetic_dataset(domain_url):
+    # 1. Fully crawl domain
+    full_knowledge = exhaustive_crawl(domain_url)
+    
+    # 2. Generate answerable queries
+    queries = llm_generate_queries(full_knowledge)
+    
+    # 3. Create query variations
+    for q in queries:
+        variations = generate_variations(q)  # synonyms, sub/super queries
+    
+    return queries, variations, full_knowledge
+```
+
+### 6.2 Evaluation Metrics
+
+1. **Efficiency**: Information gained / Pages crawled
+2. **Completeness**: Answerable queries / Total queries
+3. **Redundancy**: 1 - (Unique information / Total information)
+4. **Convergence Rate**: Pages to 95% completeness
+
+### 6.3 Ablation Studies
+
+- Impact of each score component (coverage, consistency, saturation)
+- Sensitivity to threshold parameters
+- Performance across different domain types
+
+## 7. Theoretical Properties
+
+### 7.1 Convergence Guarantee
+
+**Theorem**: For finite websites, ProgressiveCrawl converges to IS(K, Q) ≥ θ or exhausts all reachable pages.
+
+**Proof sketch**: IS(K, Q) is monotonically non-decreasing with each crawl, bounded above by 1.
+
+### 7.2 Optimality
+
+Under certain assumptions about link preview accuracy:
+- Expected crawls ≤ 2 × optimal_crawls
+- Approximation ratio improves with preview quality
+
+## 8. Implementation Design
+
+### 8.1 Core Components
+
+1. **CrawlState**: Maintains crawl history and metrics
+2. **AdaptiveConfig**: Configuration parameters
+3. **CrawlStrategy**: Pluggable strategy interface
+4. **AdaptiveCrawler**: Main orchestrator
+
+### 8.2 Integration with Crawl4AI
+
+- Wraps existing AsyncWebCrawler
+- Leverages link preview functionality
+- Maintains backward compatibility
+
+### 8.3 Persistence
+
+Knowledge base serialization for:
+- Resumable crawls
+- Knowledge sharing
+- Offline analysis
+
+## 9. Future Directions
+
+### 9.1 Advanced Scoring
+
+- Temporal information value
+- Multi-query optimization
+- Active learning from user feedback
+
+### 9.2 Distributed Crawling
+
+- Collaborative knowledge building
+- Federated information sufficiency
+
+### 9.3 Domain Adaptation
+
+- Transfer learning across domains
+- Meta-learning for threshold selection
+
+## 10. Conclusion
+
+Progressive crawling with adaptive information foraging provides a principled approach to efficient web information extraction. By combining coverage, consistency, and saturation metrics, we can determine information sufficiency without ground truth labels. The multi-strategy architecture allows graceful enhancement from pure statistical to LLM-assisted approaches based on requirements and resources.
+
+## References
+
+1. Manning, C. D., Raghavan, P., & Schütze, H. (2008). Introduction to Information Retrieval. Cambridge University Press.
+
+2. Robertson, S., & Zaragoza, H. (2009). The Probabilistic Relevance Framework: BM25 and Beyond. Foundations and Trends in Information Retrieval.
+
+3. Pirolli, P., & Card, S. (1999). Information Foraging. Psychological Review, 106(4), 643-675.
+
+4. Dasgupta, S. (2005). Analysis of a greedy active learning strategy. Advances in Neural Information Processing Systems.
+
+## Appendix A: Implementation Pseudocode
+
+```python
+class StatisticalStrategy:
+    def calculate_confidence(self, state):
+        coverage = self.calculate_coverage(state)
+        consistency = self.calculate_consistency(state)
+        saturation = self.calculate_saturation(state)
+        return min(coverage, consistency, saturation)
+    
+    def calculate_coverage(self, state):
+        # BM25-based term coverage
+        term_scores = []
+        for term in state.query.split():
+            df = state.document_frequencies.get(term, 0)
+            idf = self.idf_cache.get(term, 1.0)
+            term_scores.append(log(df + 1) * idf)
+        return mean(term_scores) / max_possible_score
+    
+    def rank_links(self, state):
+        scored_links = []
+        for link in state.pending_links:
+            relevance = self.bm25_score(link.preview_text, state.query)
+            novelty = self.calculate_novelty(link, state.knowledge_base)
+            authority = self.url_authority(link.href)
+            score = relevance * novelty * authority
+            scored_links.append((link, score))
+        return sorted(scored_links, key=lambda x: x[1], reverse=True)
+```
+
+## Appendix B: Evaluation Protocol
+
+1. **Dataset Creation**:
+   - Select diverse domains (documentation, blogs, e-commerce)
+   - Generate 100 queries per domain using LLM
+   - Create query variations (5-10 per query)
+
+2. **Baseline Comparisons**:
+   - BFS crawler (depth-limited)
+   - DFS crawler (depth-limited)
+   - Random crawler
+   - Oracle (knows relevant pages)
+
+3. **Metrics Collection**:
+   - Pages crawled vs query answerability
+   - Time to sufficient confidence
+   - False positive/negative rates
+
+4. **Statistical Analysis**:
+   - ANOVA for strategy comparison
+   - Regression for parameter sensitivity
+   - Bootstrap for confidence intervals

README.md

@@ -11,19 +11,24 @@
 [![Python Version](https://img.shields.io/pypi/pyversions/crawl4ai)](https://pypi.org/project/crawl4ai/)
 [![Downloads](https://static.pepy.tech/badge/crawl4ai/month)](https://pepy.tech/project/crawl4ai)
 
-<!-- [![Documentation Status](https://readthedocs.org/projects/crawl4ai/badge/?version=latest)](https://crawl4ai.readthedocs.io/) -->
-[![License](https://img.shields.io/github/license/unclecode/crawl4ai)](https://github.com/unclecode/crawl4ai/blob/main/LICENSE)
-[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
-[![Security: bandit](https://img.shields.io/badge/security-bandit-yellow.svg)](https://github.com/PyCQA/bandit)
-[![Contributor Covenant](https://img.shields.io/badge/Contributor%20Covenant-2.1-4baaaa.svg)](code_of_conduct.md)
-
+<p align="center">
+    <a href="https://x.com/crawl4ai">
+      <img src="https://img.shields.io/badge/Follow%20on%20X-000000?style=for-the-badge&logo=x&logoColor=white" alt="Follow on X" />
+    </a>
+    <a href="https://www.linkedin.com/company/crawl4ai">
+      <img src="https://img.shields.io/badge/Follow%20on%20LinkedIn-0077B5?style=for-the-badge&logo=linkedin&logoColor=white" alt="Follow on LinkedIn" />
+    </a>
+    <a href="https://discord.gg/jP8KfhDhyN">
+      <img src="https://img.shields.io/badge/Join%20our%20Discord-5865F2?style=for-the-badge&logo=discord&logoColor=white" alt="Join our Discord" />
+    </a>
+  </p>
 </div>
 
 Crawl4AI is the #1 trending GitHub repository, actively maintained by a vibrant community. It delivers blazing-fast, AI-ready web crawling tailored for LLMs, AI agents, and data pipelines. Open source, flexible, and built for real-time performance, Crawl4AI empowers developers with unmatched speed, precision, and deployment ease.  
 
-[✨ Check out latest update v0.6.0](#-recent-updates)
+[✨ Check out latest update v0.7.0](#-recent-updates)
 
-🎉 **Version 0.6.0 is now available!** This release candidate introduces World-aware Crawling with geolocation and locale settings, Table-to-DataFrame extraction, Browser pooling with pre-warming, Network and console traffic capture, MCP integration for AI tools, and a completely revamped Docker deployment! [Read the release notes →](https://docs.crawl4ai.com/blog)
+🎉 **Version 0.7.0 is now available!** The Adaptive Intelligence Update introduces groundbreaking features: Adaptive Crawling that learns website patterns, Virtual Scroll support for infinite pages, intelligent Link Preview with 3-layer scoring, Async URL Seeder for massive discovery, and significant performance improvements. [Read the release notes →](https://github.com/unclecode/crawl4ai/blob/main/docs/blog/release-v0.7.0.md)
 
 <details>
 <summary>🤓 <strong>My Personal Story</strong></summary>
@@ -269,8 +274,8 @@ The new Docker implementation includes:
 
 ```bash
 # Pull and run the latest release candidate
-docker pull unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number
-docker run -d -p 11235:11235 --name crawl4ai --shm-size=1g unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number
+docker pull unclecode/crawl4ai:0.7.0
+docker run -d -p 11235:11235 --name crawl4ai --shm-size=1g unclecode/crawl4ai:0.7.0
 
 # Visit the playground at http://localhost:11235/playground
 ```
@@ -291,12 +296,20 @@ import requests
 # Submit a crawl job
 response = requests.post(
     "http://localhost:11235/crawl",
-    json={"urls": "https://example.com", "priority": 10}
+    json={"urls": ["https://example.com"], "priority": 10}
 )
-task_id = response.json()["task_id"]
-
-# Continue polling until the task is complete (status="completed")
-result = requests.get(f"http://localhost:11235/task/{task_id}")
+if response.status_code == 200:
+    print("Crawl job submitted successfully.")
+    
+if "results" in response.json():
+    results = response.json()["results"]
+    print("Crawl job completed. Results:")
+    for result in results:
+        print(result)
+else:
+    task_id = response.json()["task_id"]
+    print(f"Crawl job submitted. Task ID:: {task_id}")
+    result = requests.get(f"http://localhost:11235/task/{task_id}")
 ```
 
 For more examples, see our [Docker Examples](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/docker_example.py). For advanced configuration, environment variables, and usage examples, see our [Docker Deployment Guide](https://docs.crawl4ai.com/basic/docker-deployment/).
@@ -505,7 +518,72 @@ async def test_news_crawl():
 
 ## ✨ Recent Updates
 
-### Version 0.6.0 Release Highlights
+### Version 0.7.0 Release Highlights - The Adaptive Intelligence Update
+
+- **🧠 Adaptive Crawling**: Your crawler now learns and adapts to website patterns automatically:
+  ```python
+  config = AdaptiveConfig(
+      confidence_threshold=0.7, # Min confidence to stop crawling
+      max_depth=5, # Maximum crawl depth
+      max_pages=20, # Maximum number of pages to crawl
+      strategy="statistical"
+  )
+  
+  async with AsyncWebCrawler() as crawler:
+      adaptive_crawler = AdaptiveCrawler(crawler, config)
+      state = await adaptive_crawler.digest(
+          start_url="https://news.example.com",
+          query="latest news content"
+      )
+  # Crawler learns patterns and improves extraction over time
+  ```
+
+- **🌊 Virtual Scroll Support**: Complete content extraction from infinite scroll pages:
+  ```python
+  scroll_config = VirtualScrollConfig(
+      container_selector="[data-testid='feed']",
+      scroll_count=20,
+      scroll_by="container_height",
+      wait_after_scroll=1.0
+  )
+  
+  result = await crawler.arun(url, config=CrawlerRunConfig(
+      virtual_scroll_config=scroll_config
+  ))
+  ```
+
+- **🔗 Intelligent Link Analysis**: 3-layer scoring system for smart link prioritization:
+  ```python
+  link_config = LinkPreviewConfig(
+      query="machine learning tutorials",
+      score_threshold=0.3,
+      concurrent_requests=10
+  )
+  
+  result = await crawler.arun(url, config=CrawlerRunConfig(
+      link_preview_config=link_config,
+      score_links=True
+  ))
+  # Links ranked by relevance and quality
+  ```
+
+- **🎣 Async URL Seeder**: Discover thousands of URLs in seconds:
+  ```python
+  seeder = AsyncUrlSeeder(SeedingConfig(
+      source="sitemap+cc",
+      pattern="*/blog/*",
+      query="python tutorials",
+      score_threshold=0.4
+  ))
+  
+  urls = await seeder.discover("https://example.com")
+  ```
+
+- **⚡ Performance Boost**: Up to 3x faster with optimized resource handling and memory efficiency
+
+Read the full details in our [0.7.0 Release Notes](https://docs.crawl4ai.com/blog/release-v0.7.0) or check the [CHANGELOG](https://github.com/unclecode/crawl4ai/blob/main/CHANGELOG.md).
+
+### Previous Version: 0.6.0 Release Highlights
 
 - **🌎 World-aware Crawling**: Set geolocation, language, and timezone for authentic locale-specific content:
   ```python
@@ -540,16 +618,16 @@ async def test_news_crawl():
         # Process results
         raw_df = pd.DataFrame()
         for result in results:
-            if result.success and result.media["tables"]:
+            if result.success and result.tables:
                 raw_df = pd.DataFrame(
-                    result.media["tables"][0]["rows"],
-                    columns=result.media["tables"][0]["headers"],
+                    result.tables[0]["rows"],
+                    columns=result.tables[0]["headers"],
                 )
                 break
         print(raw_df.head())
 
     finally:
-        await crawler.stop()
+        await crawler.close()
   ```
 
 - **🚀 Browser Pooling**: Pages launch hot with pre-warmed browser instances for lower latency and memory usage
@@ -575,7 +653,6 @@ async def test_news_crawl():
 
 - **📱 Multi-stage Build System**: Optimized Dockerfile with platform-specific performance enhancements
 
-Read the full details in our [0.6.0 Release Notes](https://docs.crawl4ai.com/blog/releases/0.6.0.html) or check the [CHANGELOG](https://github.com/unclecode/crawl4ai/blob/main/CHANGELOG.md).
 
 ### Previous Version: 0.5.0 Major Release Highlights
 

crawl4ai/__init__.py

@@ -3,12 +3,12 @@ import warnings
 
 from .async_webcrawler import AsyncWebCrawler, CacheMode
 # MODIFIED: Add SeedingConfig and VirtualScrollConfig here
-from .async_configs import BrowserConfig, CrawlerRunConfig, HTTPCrawlerConfig, LLMConfig, ProxyConfig, GeolocationConfig, SeedingConfig, VirtualScrollConfig
+from .async_configs import BrowserConfig, CrawlerRunConfig, HTTPCrawlerConfig, LLMConfig, ProxyConfig, GeolocationConfig, SeedingConfig, VirtualScrollConfig, LinkPreviewConfig, MatchMode
 
 from .content_scraping_strategy import (
     ContentScrapingStrategy,
-    WebScrapingStrategy,
     LXMLWebScrapingStrategy,
+    WebScrapingStrategy,  # Backward compatibility alias
 )
 from .async_logger import (
     AsyncLoggerBase,
@@ -69,6 +69,14 @@ from .deep_crawling import (
 )
 # NEW: Import AsyncUrlSeeder
 from .async_url_seeder import AsyncUrlSeeder
+# Adaptive Crawler
+from .adaptive_crawler import (
+    AdaptiveCrawler,
+    AdaptiveConfig,
+    CrawlState,
+    CrawlStrategy,
+    StatisticalStrategy
+)
 
 # C4A Script Language Support
 from .script import (
@@ -80,6 +88,13 @@ from .script import (
     ErrorDetail
 )
 
+# Browser Adapters
+from .browser_adapter import (
+    BrowserAdapter,
+    PlaywrightAdapter,
+    UndetectedAdapter
+)
+
 from .utils import (
     start_colab_display_server,
     setup_colab_environment
@@ -97,6 +112,12 @@ __all__ = [
     "VirtualScrollConfig",
     # NEW: Add AsyncUrlSeeder
     "AsyncUrlSeeder",
+    # Adaptive Crawler
+    "AdaptiveCrawler",
+    "AdaptiveConfig", 
+    "CrawlState",
+    "CrawlStrategy",
+    "StatisticalStrategy",
     "DeepCrawlStrategy",
     "BFSDeepCrawlStrategy",
     "BestFirstCrawlingStrategy",
@@ -118,6 +139,7 @@ __all__ = [
     "CrawlResult",
     "CrawlerHub",
     "CacheMode",
+    "MatchMode",
     "ContentScrapingStrategy",
     "WebScrapingStrategy",
     "LXMLWebScrapingStrategy",
@@ -159,6 +181,11 @@ __all__ = [
     "CompilationResult",
     "ValidationResult",
     "ErrorDetail",
+    # Browser Adapters
+    "BrowserAdapter",
+    "PlaywrightAdapter", 
+    "UndetectedAdapter",
+    "LinkPreviewConfig"
 ]
 
 

crawl4ai/__version__.py

@@ -1,3 +1,8 @@
-# crawl4ai/_version.py
-__version__ = "0.6.3"
+# crawl4ai/__version__.py
+
+# This is the version that will be used for stable releases
+__version__ = "0.7.2"
+
+# For nightly builds, this gets set during build process
+__nightly_version__ = None
 

crawl4ai/async_configs.py

@@ -18,17 +18,24 @@ from .extraction_strategy import ExtractionStrategy, LLMExtractionStrategy
 from .chunking_strategy import ChunkingStrategy, RegexChunking
 
 from .markdown_generation_strategy import MarkdownGenerationStrategy, DefaultMarkdownGenerator
-from .content_scraping_strategy import ContentScrapingStrategy, WebScrapingStrategy, LXMLWebScrapingStrategy
+from .content_scraping_strategy import ContentScrapingStrategy, LXMLWebScrapingStrategy
 from .deep_crawling import DeepCrawlStrategy
 
 from .cache_context import CacheMode
 from .proxy_strategy import ProxyRotationStrategy
 
-from typing import Union, List
+from typing import Union, List, Callable
 import inspect
 from typing import Any, Dict, Optional
 from enum import Enum
 
+# Type alias for URL matching
+UrlMatcher = Union[str, Callable[[str], bool], List[Union[str, Callable[[str], bool]]]]
+
+class MatchMode(Enum):
+    OR = "or"
+    AND = "and"
+
 # from .proxy_strategy import ProxyConfig
 
 
@@ -383,6 +390,8 @@ class BrowserConfig:
         light_mode (bool): Disables certain background features for performance gains. Default: False.
         extra_args (list): Additional command-line arguments passed to the browser.
                            Default: [].
+        enable_stealth (bool): If True, applies playwright-stealth to bypass basic bot detection.
+                              Cannot be used with use_undetected browser mode. Default: False.
     """
 
     def __init__(
@@ -423,6 +432,7 @@ class BrowserConfig:
         extra_args: list = None,
         debugging_port: int = 9222,
         host: str = "localhost",
+        enable_stealth: bool = False,
     ):
         self.browser_type = browser_type
         self.headless = headless 
@@ -463,6 +473,7 @@ class BrowserConfig:
         self.verbose = verbose
         self.debugging_port = debugging_port
         self.host = host
+        self.enable_stealth = enable_stealth
 
         fa_user_agenr_generator = ValidUAGenerator()
         if self.user_agent_mode == "random":
@@ -494,6 +505,13 @@ class BrowserConfig:
         # If persistent context is requested, ensure managed browser is enabled
         if self.use_persistent_context:
             self.use_managed_browser = True
+            
+        # Validate stealth configuration
+        if self.enable_stealth and self.use_managed_browser and self.browser_mode == "builtin":
+            raise ValueError(
+                "enable_stealth cannot be used with browser_mode='builtin'. "
+                "Stealth mode requires a dedicated browser instance."
+            )
 
     @staticmethod
     def from_kwargs(kwargs: dict) -> "BrowserConfig":
@@ -530,6 +548,7 @@ class BrowserConfig:
             extra_args=kwargs.get("extra_args", []),
             debugging_port=kwargs.get("debugging_port", 9222),
             host=kwargs.get("host", "localhost"),
+            enable_stealth=kwargs.get("enable_stealth", False),
         )
 
     def to_dict(self):
@@ -564,6 +583,7 @@ class BrowserConfig:
             "verbose": self.verbose,
             "debugging_port": self.debugging_port,
             "host": self.host,
+            "enable_stealth": self.enable_stealth,
         }
 
                 
@@ -862,7 +882,7 @@ class CrawlerRunConfig():
         parser_type (str): Type of parser to use for HTML parsing.
                            Default: "lxml".
         scraping_strategy (ContentScrapingStrategy): Scraping strategy to use.
-                           Default: WebScrapingStrategy.
+                           Default: LXMLWebScrapingStrategy.
         proxy_config (ProxyConfig or dict or None): Detailed proxy configuration, e.g. {"server": "...", "username": "..."}.
                                      If None, no additional proxy config. Default: None.
 
@@ -926,6 +946,8 @@ class CrawlerRunConfig():
                                Default: False.
         scroll_delay (float): Delay in seconds between scroll steps if scan_full_page is True.
                               Default: 0.2.
+        max_scroll_steps (Optional[int]): Maximum number of scroll steps to perform during full page scan.
+                                         If None, scrolls until the entire page is loaded. Default: None.
         process_iframes (bool): If True, attempts to process and inline iframe content.
                                 Default: False.
         remove_overlay_elements (bool): If True, remove overlays/popups before extracting HTML.
@@ -1066,6 +1088,7 @@ class CrawlerRunConfig():
         ignore_body_visibility: bool = True,
         scan_full_page: bool = False,
         scroll_delay: float = 0.2,
+        max_scroll_steps: Optional[int] = None,
         process_iframes: bool = False,
         remove_overlay_elements: bool = False,
         simulate_user: bool = False,
@@ -1110,6 +1133,9 @@ class CrawlerRunConfig():
         link_preview_config: Union[LinkPreviewConfig, Dict[str, Any]] = None,
         # Virtual Scroll Parameters
         virtual_scroll_config: Union[VirtualScrollConfig, Dict[str, Any]] = None,
+        # URL Matching Parameters
+        url_matcher: Optional[UrlMatcher] = None,
+        match_mode: MatchMode = MatchMode.OR,
         # Experimental Parameters
         experimental: Dict[str, Any] = None,
     ):
@@ -1170,6 +1196,7 @@ class CrawlerRunConfig():
         self.ignore_body_visibility = ignore_body_visibility
         self.scan_full_page = scan_full_page
         self.scroll_delay = scroll_delay
+        self.max_scroll_steps = max_scroll_steps
         self.process_iframes = process_iframes
         self.remove_overlay_elements = remove_overlay_elements
         self.simulate_user = simulate_user
@@ -1262,6 +1289,10 @@ class CrawlerRunConfig():
         else:
             raise ValueError("virtual_scroll_config must be VirtualScrollConfig object or dict")
         
+        # URL Matching Parameters
+        self.url_matcher = url_matcher
+        self.match_mode = match_mode
+        
         # Experimental Parameters
         self.experimental = experimental or {}
         
@@ -1317,6 +1348,51 @@ class CrawlerRunConfig():
             if "compilation error" not in str(e).lower():
                 raise ValueError(f"Failed to compile C4A script: {str(e)}")
             raise
+    
+    def is_match(self, url: str) -> bool:
+        """Check if this config matches the given URL.
+        
+        Args:
+            url: The URL to check against this config's matcher
+            
+        Returns:
+            bool: True if this config should be used for the URL or if no matcher is set.
+        """
+        if self.url_matcher is None:
+            return True
+            
+        if callable(self.url_matcher):
+            # Single function matcher
+            return self.url_matcher(url)
+        
+        elif isinstance(self.url_matcher, str):
+            # Single pattern string
+            from fnmatch import fnmatch
+            return fnmatch(url, self.url_matcher)
+        
+        elif isinstance(self.url_matcher, list):
+            # List of mixed matchers
+            if not self.url_matcher:  # Empty list
+                return False
+                
+            results = []
+            for matcher in self.url_matcher:
+                if callable(matcher):
+                    results.append(matcher(url))
+                elif isinstance(matcher, str):
+                    from fnmatch import fnmatch
+                    results.append(fnmatch(url, matcher))
+                else:
+                    # Skip invalid matchers
+                    continue
+            
+            # Apply match mode logic
+            if self.match_mode == MatchMode.OR:
+                return any(results) if results else False
+            else:  # AND mode
+                return all(results) if results else False
+        
+        return False
 
 
     def __getattr__(self, name):
@@ -1387,6 +1463,7 @@ class CrawlerRunConfig():
             ignore_body_visibility=kwargs.get("ignore_body_visibility", True),
             scan_full_page=kwargs.get("scan_full_page", False),
             scroll_delay=kwargs.get("scroll_delay", 0.2),
+            max_scroll_steps=kwargs.get("max_scroll_steps"),
             process_iframes=kwargs.get("process_iframes", False),
             remove_overlay_elements=kwargs.get("remove_overlay_elements", False),
             simulate_user=kwargs.get("simulate_user", False),
@@ -1438,6 +1515,9 @@ class CrawlerRunConfig():
             # Link Extraction Parameters
             link_preview_config=kwargs.get("link_preview_config"),
             url=kwargs.get("url"),
+            # URL Matching Parameters
+            url_matcher=kwargs.get("url_matcher"),
+            match_mode=kwargs.get("match_mode", MatchMode.OR),
             # Experimental Parameters 
             experimental=kwargs.get("experimental"),
         )
@@ -1499,6 +1579,7 @@ class CrawlerRunConfig():
             "ignore_body_visibility": self.ignore_body_visibility,
             "scan_full_page": self.scan_full_page,
             "scroll_delay": self.scroll_delay,
+            "max_scroll_steps": self.max_scroll_steps,
             "process_iframes": self.process_iframes,
             "remove_overlay_elements": self.remove_overlay_elements,
             "simulate_user": self.simulate_user,
@@ -1534,6 +1615,8 @@ class CrawlerRunConfig():
             "deep_crawl_strategy": self.deep_crawl_strategy,
             "link_preview_config": self.link_preview_config.to_dict() if self.link_preview_config else None,
             "url": self.url,
+            "url_matcher": self.url_matcher,
+            "match_mode": self.match_mode,
             "experimental": self.experimental,
         }
 
@@ -1653,22 +1736,57 @@ class SeedingConfig:
     """
     def __init__(
         self,
-        source: str = "sitemap+cc",  # Options: "sitemap", "cc", "sitemap+cc"
-        pattern: Optional[str] = "*", # URL pattern to filter discovered URLs (e.g., "*example.com/blog/*")
-        live_check: bool = False,    # Whether to perform HEAD requests to verify URL liveness
-        extract_head: bool = False,  # Whether to fetch and parse <head> section for metadata
-        max_urls: int = -1, # Maximum number of URLs to discover (default: -1 for no limit)
-        concurrency: int = 1000,      # Maximum concurrent requests for live checks/head extraction
-        hits_per_sec: int = 5,      # Rate limit in requests per second
-        force: bool = False, # If True, bypasses the AsyncUrlSeeder's internal .jsonl cache
-        base_directory: Optional[str] = None, # Base directory for UrlSeeder's cache files (.jsonl)
-        llm_config: Optional[LLMConfig] = None, # Forward LLM config for future use (e.g., relevance scoring)
-        verbose: Optional[bool] = None, # Override crawler's general verbose setting
-        query: Optional[str] = None,  # Search query for relevance scoring
-        score_threshold: Optional[float] = None,  # Minimum relevance score to include URL (0.0-1.0)
-        scoring_method: str = "bm25",  # Scoring method: "bm25" (default), future: "semantic"
-        filter_nonsense_urls: bool = True,  # Filter out utility URLs like robots.txt, sitemap.xml, etc.
+        source: str = "sitemap+cc",
+        pattern: Optional[str] = "*",
+        live_check: bool = False,
+        extract_head: bool = False,
+        max_urls: int = -1,
+        concurrency: int = 1000,
+        hits_per_sec: int = 5,
+        force: bool = False,
+        base_directory: Optional[str] = None,
+        llm_config: Optional[LLMConfig] = None,
+        verbose: Optional[bool] = None,
+        query: Optional[str] = None,
+        score_threshold: Optional[float] = None,
+        scoring_method: str = "bm25",
+        filter_nonsense_urls: bool = True,
     ):
+        """
+        Initialize URL seeding configuration.
+        
+        Args:
+            source: Discovery source(s) to use. Options: "sitemap", "cc" (Common Crawl), 
+                   or "sitemap+cc" (both). Default: "sitemap+cc"
+            pattern: URL pattern to filter discovered URLs (e.g., "*example.com/blog/*"). 
+                    Supports glob-style wildcards. Default: "*" (all URLs)
+            live_check: Whether to perform HEAD requests to verify URL liveness. 
+                       Default: False
+            extract_head: Whether to fetch and parse <head> section for metadata extraction.
+                         Required for BM25 relevance scoring. Default: False
+            max_urls: Maximum number of URLs to discover. Use -1 for no limit. 
+                     Default: -1
+            concurrency: Maximum concurrent requests for live checks/head extraction. 
+                        Default: 1000
+            hits_per_sec: Rate limit in requests per second to avoid overwhelming servers. 
+                         Default: 5
+            force: If True, bypasses the AsyncUrlSeeder's internal .jsonl cache and 
+                  re-fetches URLs. Default: False
+            base_directory: Base directory for UrlSeeder's cache files (.jsonl). 
+                           If None, uses default ~/.crawl4ai/. Default: None
+            llm_config: LLM configuration for future features (e.g., semantic scoring). 
+                       Currently unused. Default: None
+            verbose: Override crawler's general verbose setting for seeding operations. 
+                    Default: None (inherits from crawler)
+            query: Search query for BM25 relevance scoring (e.g., "python tutorials"). 
+                  Requires extract_head=True. Default: None
+            score_threshold: Minimum relevance score (0.0-1.0) to include URL. 
+                           Only applies when query is provided. Default: None
+            scoring_method: Scoring algorithm to use. Currently only "bm25" is supported. 
+                          Future: "semantic". Default: "bm25"
+            filter_nonsense_urls: Filter out utility URLs like robots.txt, sitemap.xml, 
+                                 ads.txt, favicon.ico, etc. Default: True
+        """
         self.source = source
         self.pattern = pattern
         self.live_check = live_check

crawl4ai/async_crawler_strategy.py

@@ -21,6 +21,7 @@ from .async_logger import AsyncLogger
 from .ssl_certificate import SSLCertificate
 from .user_agent_generator import ValidUAGenerator
 from .browser_manager import BrowserManager
+from .browser_adapter import BrowserAdapter, PlaywrightAdapter, UndetectedAdapter
 
 import aiofiles
 import aiohttp
@@ -71,7 +72,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
     """
 
     def __init__(
-        self, browser_config: BrowserConfig = None, logger: AsyncLogger = None, **kwargs
+        self, browser_config: BrowserConfig = None, logger: AsyncLogger = None, browser_adapter: BrowserAdapter = None, **kwargs
     ):
         """
         Initialize the AsyncPlaywrightCrawlerStrategy with a browser configuration.
@@ -80,11 +81,16 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             browser_config (BrowserConfig): Configuration object containing browser settings.
                                           If None, will be created from kwargs for backwards compatibility.
             logger: Logger instance for recording events and errors.
+            browser_adapter (BrowserAdapter): Browser adapter for handling browser-specific operations.
+                                           If None, defaults to PlaywrightAdapter.
             **kwargs: Additional arguments for backwards compatibility and extending functionality.
         """
         # Initialize browser config, either from provided object or kwargs
         self.browser_config = browser_config or BrowserConfig.from_kwargs(kwargs)
         self.logger = logger
+        
+        # Initialize browser adapter
+        self.adapter = browser_adapter or PlaywrightAdapter()
 
         # Initialize session management
         self._downloaded_files = []
@@ -104,7 +110,9 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
 
         # Initialize browser manager with config
         self.browser_manager = BrowserManager(
-            browser_config=self.browser_config, logger=self.logger
+            browser_config=self.browser_config, 
+            logger=self.logger,
+            use_undetected=isinstance(self.adapter, UndetectedAdapter)
         )
 
     async def __aenter__(self):
@@ -322,7 +330,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         """
 
         try:
-            result = await page.evaluate(wrapper_js)
+            result = await self.adapter.evaluate(page, wrapper_js)
             return result
         except Exception as e:
             if "Error evaluating condition" in str(e):
@@ -367,7 +375,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
 
                     # Replace the iframe with a div containing the extracted content
                     _iframe = iframe_content.replace("`", "\\`")
-                    await page.evaluate(
+                    await self.adapter.evaluate(page,
                         f"""
                         () => {{
                             const iframe = document.getElementById('iframe-{i}');
@@ -445,6 +453,9 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             return await self._crawl_web(url, config)
 
         elif url.startswith("file://"):
+            # initialize empty lists for console messages
+            captured_console = []
+            
             # Process local file
             local_file_path = url[7:]  # Remove 'file://' prefix
             if not os.path.exists(local_file_path):
@@ -466,9 +477,15 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                 console_messages=captured_console,
             )
 
-        elif url.startswith("raw:") or url.startswith("raw://"):
+        ##### 
+        # Since both "raw:" and "raw://" start with "raw:", the first condition is always true for both, so "raw://" will be sliced as "//...", which is incorrect.
+        # Fix: Check for "raw://" first, then "raw:"
+        # Also, the prefix "raw://" is actually 6 characters long, not 7, so it should be sliced accordingly: url[6:]
+        #####
+        elif url.startswith("raw://") or url.startswith("raw:"):
             # Process raw HTML content
-            raw_html = url[4:] if url[:4] == "raw:" else url[7:]
+            # raw_html = url[4:] if url[:4] == "raw:" else url[7:]
+            raw_html = url[6:] if url.startswith("raw://") else url[4:]
             html = raw_html
             if config.screenshot:
                 screenshot_data = await self._generate_screenshot_from_html(html)
@@ -619,91 +636,16 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             page.on("requestfailed", handle_request_failed_capture)
 
         # Console Message Capturing
+        handle_console = None
+        handle_error = None
         if config.capture_console_messages:
-            def handle_console_capture(msg):
-                try:
-                    message_type = "unknown"
-                    try:
-                        message_type = msg.type
-                    except:
-                        pass
-                        
-                    message_text = "unknown"
-                    try:
-                        message_text = msg.text
-                    except:
-                        pass
-                        
-                    # Basic console message with minimal content
-                    entry = {
-                        "type": message_type,
-                        "text": message_text,
-                        "timestamp": time.time()
-                    }
-                    
-                    captured_console.append(entry)
-                    
-                except Exception as e:
-                    if self.logger:
-                        self.logger.warning(f"Error capturing console message: {e}", tag="CAPTURE")
-                    # Still add something to the list even on error
-                    captured_console.append({
-                        "type": "console_capture_error", 
-                        "error": str(e), 
-                        "timestamp": time.time()
-                    })
-
-            def handle_pageerror_capture(err):
-                try:
-                    error_message = "Unknown error"
-                    try:
-                        error_message = err.message
-                    except:
-                        pass
-                        
-                    error_stack = ""
-                    try:
-                        error_stack = err.stack
-                    except:
-                        pass
-                        
-                    captured_console.append({
-                        "type": "error",
-                        "text": error_message,
-                        "stack": error_stack,
-                        "timestamp": time.time()
-                    })
-                except Exception as e:
-                    if self.logger:
-                        self.logger.warning(f"Error capturing page error: {e}", tag="CAPTURE")
-                    captured_console.append({
-                        "type": "pageerror_capture_error", 
-                        "error": str(e), 
-                        "timestamp": time.time()
-                    })
-
-            # Add event listeners directly
-            page.on("console", handle_console_capture)
-            page.on("pageerror", handle_pageerror_capture)
+            # Set up console capture using adapter
+            handle_console = await self.adapter.setup_console_capture(page, captured_console)
+            handle_error = await self.adapter.setup_error_capture(page, captured_console)
 
         # Set up console logging if requested
-        if config.log_console:
-            def log_consol(
-                msg, console_log_type="debug"
-            ):  # Corrected the parameter syntax
-                if console_log_type == "error":
-                    self.logger.error(
-                        message=f"Console error: {msg}",  # Use f-string for variable interpolation
-                        tag="CONSOLE"
-                    )
-                elif console_log_type == "debug":
-                    self.logger.debug(
-                        message=f"Console: {msg}",  # Use f-string for variable interpolation
-                        tag="CONSOLE"
-                    )
-
-            page.on("console", log_consol)
-            page.on("pageerror", lambda e: log_consol(e, "error"))
+        # Note: For undetected browsers, console logging won't work directly
+        # but captured messages can still be logged after retrieval
 
         try:
             # Get SSL certificate information if requested and URL is HTTPS
@@ -741,18 +683,49 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                     )
                     redirected_url = page.url
                 except Error as e:
-                    raise RuntimeError(f"Failed on navigating ACS-GOTO:\n{str(e)}")
+                    # Allow navigation to be aborted when downloading files
+                    # This is expected behavior for downloads in some browser engines
+                    if 'net::ERR_ABORTED' in str(e) and self.browser_config.accept_downloads:
+                        self.logger.info(
+                            message=f"Navigation aborted, likely due to file download: {url}",
+                            tag="GOTO",
+                            params={"url": url},
+                        )
+                        response = None
+                    else:
+                        raise RuntimeError(f"Failed on navigating ACS-GOTO:\n{str(e)}")
 
                 await self.execute_hook(
                     "after_goto", page, context=context, url=url, response=response, config=config
                 )
 
+                # ──────────────────────────────────────────────────────────────
+                # Walk the redirect chain.  Playwright returns only the last
+                # hop, so we trace the `request.redirected_from` links until the
+                # first response that differs from the final one and surface its
+                # status-code.
+                # ──────────────────────────────────────────────────────────────
                 if response is None:
                     status_code = 200
                     response_headers = {}
                 else:
-                    status_code = response.status
-                    response_headers = response.headers
+                    first_resp = response
+                    req = response.request
+                    while req and req.redirected_from:
+                        prev_req = req.redirected_from
+                        prev_resp = await prev_req.response()
+                        if prev_resp:                       # keep earliest
+                            first_resp = prev_resp
+                        req = prev_req
+                
+                    status_code = first_resp.status
+                    response_headers = first_resp.headers
+                # if response is None:
+                #     status_code = 200
+                #     response_headers = {}
+                # else:
+                #     status_code = response.status
+                #     response_headers = response.headers
 
             else:
                 status_code = 200
@@ -784,7 +757,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             except Error:
                 visibility_info = await self.check_visibility(page)
 
-                if self.browser_config.config.verbose:
+                if self.browser_config.verbose:
                     self.logger.debug(
                         message="Body visibility info: {info}",
                         tag="DEBUG",
@@ -896,7 +869,8 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
 
             # Handle full page scanning
             if config.scan_full_page:
-                await self._handle_full_page_scan(page, config.scroll_delay)
+                # await self._handle_full_page_scan(page, config.scroll_delay)
+                await self._handle_full_page_scan(page, config.scroll_delay, config.max_scroll_steps)
 
             # Handle virtual scroll if configured
             if config.virtual_scroll_config:
@@ -957,7 +931,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                         await page.wait_for_load_state("domcontentloaded", timeout=5)
                     except PlaywrightTimeoutError:
                         pass
-                    await page.evaluate(update_image_dimensions_js)
+                    await self.adapter.evaluate(page, update_image_dimensions_js)
                 except Exception as e:
                     self.logger.error(
                         message="Error updating image dimensions: {error}",
@@ -986,7 +960,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                     
                     for selector in selectors:
                         try:
-                            content = await page.evaluate(
+                            content = await self.adapter.evaluate(page,
                                 f"""Array.from(document.querySelectorAll("{selector}"))
                                     .map(el => el.outerHTML)
                                     .join('')"""
@@ -1044,6 +1018,11 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                 await asyncio.sleep(delay)
                 return await page.content()
 
+            # For undetected browsers, retrieve console messages before returning
+            if config.capture_console_messages and hasattr(self.adapter, 'retrieve_console_messages'):
+                final_messages = await self.adapter.retrieve_console_messages(page)
+                captured_console.extend(final_messages)
+
             # Return complete response
             return AsyncCrawlResponse(
                 html=html,
@@ -1082,13 +1061,19 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                     page.remove_listener("response", handle_response_capture)
                     page.remove_listener("requestfailed", handle_request_failed_capture)
                 if config.capture_console_messages:
-                    page.remove_listener("console", handle_console_capture)
-                    page.remove_listener("pageerror", handle_pageerror_capture)
+                    # Retrieve any final console messages for undetected browsers
+                    if hasattr(self.adapter, 'retrieve_console_messages'):
+                        final_messages = await self.adapter.retrieve_console_messages(page)
+                        captured_console.extend(final_messages)
+                    
+                    # Clean up console capture
+                    await self.adapter.cleanup_console_capture(page, handle_console, handle_error)
                 
                 # Close the page
                 await page.close()
 
-    async def _handle_full_page_scan(self, page: Page, scroll_delay: float = 0.1):
+    # async def _handle_full_page_scan(self, page: Page, scroll_delay: float = 0.1):
+    async def _handle_full_page_scan(self, page: Page, scroll_delay: float = 0.1, max_scroll_steps: Optional[int] = None):
         """
         Helper method to handle full page scanning.
 
@@ -1103,6 +1088,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         Args:
             page (Page): The Playwright page object
             scroll_delay (float): The delay between page scrolls
+            max_scroll_steps (Optional[int]): Maximum number of scroll steps to perform. If None, scrolls until end.
 
         """
         try:
@@ -1127,9 +1113,21 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             dimensions = await self.get_page_dimensions(page)
             total_height = dimensions["height"]
 
+            scroll_step_count = 0
             while current_position < total_height:
+                #### 
+                # NEW FEATURE: Check if we've reached the maximum allowed scroll steps
+                # This prevents infinite scrolling on very long pages or infinite scroll scenarios
+                # If max_scroll_steps is None, this check is skipped (unlimited scrolling - original behavior)
+                ####
+                if max_scroll_steps is not None and scroll_step_count >= max_scroll_steps:
+                    break
                 current_position = min(current_position + viewport_height, total_height)
                 await self.safe_scroll(page, 0, current_position, delay=scroll_delay)
+
+                # Increment the step counter for max_scroll_steps tracking
+                scroll_step_count += 1
+                
                 # await page.evaluate(f"window.scrollTo(0, {current_position})")
                 # await asyncio.sleep(scroll_delay)
 
@@ -1299,7 +1297,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             """
             
             # Execute virtual scroll capture
-            result = await page.evaluate(virtual_scroll_js, config.to_dict())
+            result = await self.adapter.evaluate(page, virtual_scroll_js, config.to_dict())
             
             if result.get("replaced", False):
                 self.logger.success(
@@ -1383,7 +1381,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         remove_overlays_js = load_js_script("remove_overlay_elements")
 
         try:
-            await page.evaluate(
+            await self.adapter.evaluate(page,
                 f"""
                 (() => {{
                     try {{
@@ -1616,12 +1614,32 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             num_segments = (page_height // viewport_height) + 1
             for i in range(num_segments):
                 y_offset = i * viewport_height
+                # Special handling for the last segment
+                if i == num_segments - 1:
+                    last_part_height = page_height % viewport_height
+                    
+                    # If page_height is an exact multiple of viewport_height,
+                    # we don't need an extra segment
+                    if last_part_height == 0:
+                        # Skip last segment if page height is exact multiple of viewport
+                        break
+                    
+                    # Adjust viewport to exactly match the remaining content height
+                    await page.set_viewport_size({"width": page_width, "height": last_part_height})
+                
                 await page.evaluate(f"window.scrollTo(0, {y_offset})")
                 await asyncio.sleep(0.01)  # wait for render
-                seg_shot = await page.screenshot(full_page=False)
+                
+                # Capture the current segment
+                # Note: Using compression options (format, quality) would go here
+                seg_shot = await page.screenshot(full_page=False, type="jpeg", quality=85)
+                # seg_shot = await page.screenshot(full_page=False)
                 img = Image.open(BytesIO(seg_shot)).convert("RGB")
                 segments.append(img)
 
+            # Reset viewport to original size after capturing segments
+            await page.set_viewport_size({"width": page_width, "height": viewport_height})
+
             total_height = sum(img.height for img in segments)
             stitched = Image.new("RGB", (segments[0].width, total_height))
             offset = 0
@@ -1750,12 +1768,31 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                     # then wait for the new page to load before continuing
                     result = None
                     try:
-                        result = await page.evaluate(
+                        # OLD VERSION:
+                        # result = await page.evaluate(
+                        #     f"""
+                        # (async () => {{
+                        #     try {{
+                        #         const script_result = {script};
+                        #         return {{ success: true, result: script_result }};
+                        #     }} catch (err) {{
+                        #         return {{ success: false, error: err.toString(), stack: err.stack }};
+                        #     }}
+                        # }})();
+                        # """
+                        # )
+                        
+                        # """ NEW VERSION:
+                        # When {script} contains statements (e.g., const link = …; link.click();), 
+                        # this forms invalid JavaScript, causing Playwright execution error: SyntaxError: Unexpected token 'const'.
+                        # """
+                        result = await self.adapter.evaluate(page,
                             f"""
                         (async () => {{
                             try {{
-                                const script_result = {script};
-                                return {{ success: true, result: script_result }};
+                                return await (async () => {{
+                                    {script}
+                                }})();
                             }} catch (err) {{
                                 return {{ success: false, error: err.toString(), stack: err.stack }};
                             }}
@@ -1871,7 +1908,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             for script in scripts:
                 try:
                     # Execute the script and wait for network idle
-                    result = await page.evaluate(
+                    result = await self.adapter.evaluate(page,
                         f"""
                         (() => {{
                             return new Promise((resolve) => {{
@@ -1955,7 +1992,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         Returns:
             Boolean indicating visibility
         """
-        return await page.evaluate(
+        return await self.adapter.evaluate(page,
             """
             () => {
                 const element = document.body;
@@ -1996,7 +2033,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             Dict containing scroll status and position information
         """
         try:
-            result = await page.evaluate(
+            result = await self.adapter.evaluate(page,
                 f"""() => {{
                     try {{
                         const startX = window.scrollX;
@@ -2053,7 +2090,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         Returns:
             Dict containing width and height of the page
         """
-        return await page.evaluate(
+        return await self.adapter.evaluate(page,
             """
             () => {
                 const {scrollWidth, scrollHeight} = document.documentElement;
@@ -2073,7 +2110,7 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             bool: True if page needs scrolling
         """
         try:
-            need_scroll = await page.evaluate(
+            need_scroll = await self.adapter.evaluate(page,
                 """
             () => {
                 const scrollHeight = document.documentElement.scrollHeight;
@@ -2092,265 +2129,3 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             return True  # Default to scrolling if check fails
 
 
-####################################################################################################
-# HTTP Crawler Strategy
-####################################################################################################
-
-class HTTPCrawlerError(Exception):
-    """Base error class for HTTP crawler specific exceptions"""
-    pass
-
-
-class ConnectionTimeoutError(HTTPCrawlerError):
-    """Raised when connection timeout occurs"""
-    pass
-
-
-class HTTPStatusError(HTTPCrawlerError):
-    """Raised for unexpected status codes"""
-    def __init__(self, status_code: int, message: str):
-        self.status_code = status_code
-        super().__init__(f"HTTP {status_code}: {message}")
-
-
-class AsyncHTTPCrawlerStrategy(AsyncCrawlerStrategy):
-    """
-    Fast, lightweight HTTP-only crawler strategy optimized for memory efficiency.
-    """
-    
-    __slots__ = ('logger', 'max_connections', 'dns_cache_ttl', 'chunk_size', '_session', 'hooks', 'browser_config')
-
-    DEFAULT_TIMEOUT: Final[int] = 30
-    DEFAULT_CHUNK_SIZE: Final[int] = 64 * 1024  
-    DEFAULT_MAX_CONNECTIONS: Final[int] = min(32, (os.cpu_count() or 1) * 4)
-    DEFAULT_DNS_CACHE_TTL: Final[int] = 300
-    VALID_SCHEMES: Final = frozenset({'http', 'https', 'file', 'raw'})
-
-    _BASE_HEADERS: Final = MappingProxyType({
-        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
-        'Accept-Language': 'en-US,en;q=0.5',
-        'Accept-Encoding': 'gzip, deflate, br',
-        'Connection': 'keep-alive',
-        'Upgrade-Insecure-Requests': '1',
-        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
-    })
-    
-    def __init__(
-        self, 
-        browser_config: Optional[HTTPCrawlerConfig] = None,
-        logger: Optional[AsyncLogger] = None,
-        max_connections: int = DEFAULT_MAX_CONNECTIONS,
-        dns_cache_ttl: int = DEFAULT_DNS_CACHE_TTL,
-        chunk_size: int = DEFAULT_CHUNK_SIZE
-    ):
-        """Initialize the HTTP crawler with config"""
-        self.browser_config = browser_config or HTTPCrawlerConfig()
-        self.logger = logger
-        self.max_connections = max_connections
-        self.dns_cache_ttl = dns_cache_ttl
-        self.chunk_size = chunk_size
-        self._session: Optional[aiohttp.ClientSession] = None
-        
-        self.hooks = {
-            k: partial(self._execute_hook, k) 
-            for k in ('before_request', 'after_request', 'on_error')
-        }
-
-        # Set default hooks
-        self.set_hook('before_request', lambda *args, **kwargs: None)
-        self.set_hook('after_request', lambda *args, **kwargs: None)
-        self.set_hook('on_error', lambda *args, **kwargs: None)
-                      
-
-    async def __aenter__(self) -> AsyncHTTPCrawlerStrategy:
-        await self.start()
-        return self
-        
-    async def __aexit__(self, exc_type, exc_val, exc_tb) -> None:
-        await self.close()
-
-    @contextlib.asynccontextmanager
-    async def _session_context(self):
-        try:
-            if not self._session:
-                await self.start()
-            yield self._session
-        finally:
-            pass
-
-    def set_hook(self, hook_type: str, hook_func: Callable) -> None:
-        if hook_type in self.hooks:
-            self.hooks[hook_type] = partial(self._execute_hook, hook_type, hook_func)
-        else:
-            raise ValueError(f"Invalid hook type: {hook_type}")
-
-    async def _execute_hook(
-        self, 
-        hook_type: str, 
-        hook_func: Callable,
-        *args: Any, 
-        **kwargs: Any
-    ) -> Any:
-        if asyncio.iscoroutinefunction(hook_func):
-            return await hook_func(*args, **kwargs)
-        return hook_func(*args, **kwargs)
-
-    async def start(self) -> None:
-        if not self._session:
-            connector = aiohttp.TCPConnector(
-                limit=self.max_connections,
-                ttl_dns_cache=self.dns_cache_ttl,
-                use_dns_cache=True,
-                force_close=False
-            )
-            self._session = aiohttp.ClientSession(
-                headers=dict(self._BASE_HEADERS),
-                connector=connector,
-                timeout=ClientTimeout(total=self.DEFAULT_TIMEOUT)
-            )
-
-    async def close(self) -> None:
-        if self._session and not self._session.closed:
-            try:
-                await asyncio.wait_for(self._session.close(), timeout=5.0)
-            except asyncio.TimeoutError:
-                if self.logger:
-                    self.logger.warning(
-                        message="Session cleanup timed out",
-                        tag="CLEANUP"
-                    )
-            finally:
-                self._session = None
-
-    async def _stream_file(self, path: str) -> AsyncGenerator[memoryview, None]:
-        async with aiofiles.open(path, mode='rb') as f:
-            while chunk := await f.read(self.chunk_size):
-                yield memoryview(chunk)
-
-    async def _handle_file(self, path: str) -> AsyncCrawlResponse:
-        if not os.path.exists(path):
-            raise FileNotFoundError(f"Local file not found: {path}")
-            
-        chunks = []
-        async for chunk in self._stream_file(path):
-            chunks.append(chunk.tobytes().decode('utf-8', errors='replace'))
-            
-        return AsyncCrawlResponse(
-            html=''.join(chunks),
-            response_headers={},
-            status_code=200
-        )
-
-    async def _handle_raw(self, content: str) -> AsyncCrawlResponse:
-        return AsyncCrawlResponse(
-            html=content,
-            response_headers={},
-            status_code=200
-        )
-
-
-    async def _handle_http(
-        self, 
-        url: str, 
-        config: CrawlerRunConfig
-    ) -> AsyncCrawlResponse:
-        async with self._session_context() as session:
-            timeout = ClientTimeout(
-                total=config.page_timeout or self.DEFAULT_TIMEOUT,
-                connect=10,
-                sock_read=30
-            )
-            
-            headers = dict(self._BASE_HEADERS)
-            if self.browser_config.headers:
-                headers.update(self.browser_config.headers)
-
-            request_kwargs = {
-                'timeout': timeout,
-                'allow_redirects': self.browser_config.follow_redirects,
-                'ssl': self.browser_config.verify_ssl,
-                'headers': headers
-            }
-
-            if self.browser_config.method == "POST":
-                if self.browser_config.data:
-                    request_kwargs['data'] = self.browser_config.data
-                if self.browser_config.json:
-                    request_kwargs['json'] = self.browser_config.json
-
-            await self.hooks['before_request'](url, request_kwargs)
-
-            try:
-                async with session.request(self.browser_config.method, url, **request_kwargs) as response:
-                    content = memoryview(await response.read())
-                    
-                    if not (200 <= response.status < 300):
-                        raise HTTPStatusError(
-                            response.status,
-                            f"Unexpected status code for {url}"
-                        )
-                    
-                    encoding = response.charset
-                    if not encoding:
-                        encoding = chardet.detect(content.tobytes())['encoding'] or 'utf-8'                    
-                    
-                    result = AsyncCrawlResponse(
-                        html=content.tobytes().decode(encoding, errors='replace'),
-                        response_headers=dict(response.headers),
-                        status_code=response.status,
-                        redirected_url=str(response.url)
-                    )
-                    
-                    await self.hooks['after_request'](result)
-                    return result
-
-            except aiohttp.ServerTimeoutError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionTimeoutError(f"Request timed out: {str(e)}")
-                
-            except aiohttp.ClientConnectorError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionError(f"Connection failed: {str(e)}")
-                
-            except aiohttp.ClientError as e:
-                await self.hooks['on_error'](e)
-                raise HTTPCrawlerError(f"HTTP client error: {str(e)}")
-            
-            except asyncio.exceptions.TimeoutError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionTimeoutError(f"Request timed out: {str(e)}")
-            
-            except Exception as e:
-                await self.hooks['on_error'](e)
-                raise HTTPCrawlerError(f"HTTP request failed: {str(e)}")
-
-    async def crawl(
-        self, 
-        url: str, 
-        config: Optional[CrawlerRunConfig] = None, 
-        **kwargs
-    ) -> AsyncCrawlResponse:
-        config = config or CrawlerRunConfig.from_kwargs(kwargs)
-        
-        parsed = urlparse(url)
-        scheme = parsed.scheme.rstrip('/')
-        
-        if scheme not in self.VALID_SCHEMES:
-            raise ValueError(f"Unsupported URL scheme: {scheme}")
-            
-        try:
-            if scheme == 'file':
-                return await self._handle_file(parsed.path)
-            elif scheme == 'raw':
-                return await self._handle_raw(parsed.path)
-            else:  # http or https
-                return await self._handle_http(url, config)
-                
-        except Exception as e:
-            if self.logger:
-                self.logger.error(
-                    message="Crawl failed: {error}",
-                    tag="CRAWL",
-                    params={"error": str(e), "url": url}
-                )
-            raise

crawl4ai/async_dispatcher.py

@@ -1,4 +1,4 @@
-from typing import Dict, Optional, List, Tuple
+from typing import Dict, Optional, List, Tuple, Union
 from .async_configs import CrawlerRunConfig
 from .models import (
     CrawlResult,
@@ -22,6 +22,8 @@ from urllib.parse import urlparse
 import random
 from abc import ABC, abstractmethod
 
+from .memory_utils import get_true_memory_usage_percent
+
 
 class RateLimiter:
     def __init__(
@@ -96,11 +98,37 @@ class BaseDispatcher(ABC):
         self.rate_limiter = rate_limiter
         self.monitor = monitor
 
+    def select_config(self, url: str, configs: Union[CrawlerRunConfig, List[CrawlerRunConfig]]) -> Optional[CrawlerRunConfig]:
+        """Select the appropriate config for a given URL.
+        
+        Args:
+            url: The URL to match against
+            configs: Single config or list of configs to choose from
+            
+        Returns:
+            The matching config, or None if no match found
+        """
+        # Single config - return as is
+        if isinstance(configs, CrawlerRunConfig):
+            return configs
+        
+        # Empty list - return None
+        if not configs:
+            return None
+        
+        # Find first matching config
+        for config in configs:
+            if config.is_match(url):
+                return config
+        
+        # No match found - return None to indicate URL should be skipped
+        return None
+
     @abstractmethod
     async def crawl_url(
         self,
         url: str,
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
         task_id: str,
         monitor: Optional[CrawlerMonitor] = None,
     ) -> CrawlerTaskResult:
@@ -111,7 +139,7 @@ class BaseDispatcher(ABC):
         self,
         urls: List[str],
         crawler: AsyncWebCrawler,  # noqa: F821
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
         monitor: Optional[CrawlerMonitor] = None,
     ) -> List[CrawlerTaskResult]:
         pass
@@ -147,7 +175,7 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
     async def _memory_monitor_task(self):
         """Background task to continuously monitor memory usage and update state"""
         while True:
-            self.current_memory_percent = psutil.virtual_memory().percent
+            self.current_memory_percent = get_true_memory_usage_percent()
 
             # Enter memory pressure mode if we cross the threshold
             if self.current_memory_percent >= self.memory_threshold_percent:
@@ -200,7 +228,7 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
     async def crawl_url(
         self,
         url: str,
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
         task_id: str,
         retry_count: int = 0,
     ) -> CrawlerTaskResult:
@@ -208,6 +236,37 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
         error_message = ""
         memory_usage = peak_memory = 0.0
         
+        # Select appropriate config for this URL
+        selected_config = self.select_config(url, config)
+        
+        # If no config matches, return failed result
+        if selected_config is None:
+            error_message = f"No matching configuration found for URL: {url}"
+            if self.monitor:
+                self.monitor.update_task(
+                    task_id, 
+                    status=CrawlStatus.FAILED,
+                    error_message=error_message
+                )
+            
+            return CrawlerTaskResult(
+                task_id=task_id,
+                url=url,
+                result=CrawlResult(
+                    url=url, 
+                    html="", 
+                    metadata={"status": "no_config_match"}, 
+                    success=False, 
+                    error_message=error_message
+                ),
+                memory_usage=0,
+                peak_memory=0,
+                start_time=start_time,
+                end_time=time.time(),
+                error_message=error_message,
+                retry_count=retry_count
+            )
+        
         # Get starting memory for accurate measurement
         process = psutil.Process()
         start_memory = process.memory_info().rss / (1024 * 1024)
@@ -257,8 +316,8 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
                     retry_count=retry_count + 1
                 )
             
-            # Execute the crawl
-            result = await self.crawler.arun(url, config=config, session_id=task_id)
+            # Execute the crawl with selected config
+            result = await self.crawler.arun(url, config=selected_config, session_id=task_id)
             
             # Measure memory usage
             end_memory = process.memory_info().rss / (1024 * 1024)
@@ -316,7 +375,7 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
         self,
         urls: List[str],
         crawler: AsyncWebCrawler,
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
     ) -> List[CrawlerTaskResult]:
         self.crawler = crawler
         
@@ -470,7 +529,7 @@ class MemoryAdaptiveDispatcher(BaseDispatcher):
         self,
         urls: List[str],
         crawler: AsyncWebCrawler,
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
     ) -> AsyncGenerator[CrawlerTaskResult, None]:
         self.crawler = crawler
         
@@ -572,7 +631,7 @@ class SemaphoreDispatcher(BaseDispatcher):
     async def crawl_url(
         self,
         url: str,
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
         task_id: str,
         semaphore: asyncio.Semaphore = None,
     ) -> CrawlerTaskResult:
@@ -580,6 +639,36 @@ class SemaphoreDispatcher(BaseDispatcher):
         error_message = ""
         memory_usage = peak_memory = 0.0
 
+        # Select appropriate config for this URL
+        selected_config = self.select_config(url, config)
+        
+        # If no config matches, return failed result
+        if selected_config is None:
+            error_message = f"No matching configuration found for URL: {url}"
+            if self.monitor:
+                self.monitor.update_task(
+                    task_id, 
+                    status=CrawlStatus.FAILED,
+                    error_message=error_message
+                )
+            
+            return CrawlerTaskResult(
+                task_id=task_id,
+                url=url,
+                result=CrawlResult(
+                    url=url, 
+                    html="", 
+                    metadata={"status": "no_config_match"}, 
+                    success=False, 
+                    error_message=error_message
+                ),
+                memory_usage=0,
+                peak_memory=0,
+                start_time=start_time,
+                end_time=time.time(),
+                error_message=error_message
+            )
+
         try:
             if self.monitor:
                 self.monitor.update_task(
@@ -592,7 +681,7 @@ class SemaphoreDispatcher(BaseDispatcher):
             async with semaphore:
                 process = psutil.Process()
                 start_memory = process.memory_info().rss / (1024 * 1024)
-                result = await self.crawler.arun(url, config=config, session_id=task_id)
+                result = await self.crawler.arun(url, config=selected_config, session_id=task_id)
                 end_memory = process.memory_info().rss / (1024 * 1024)
 
                 memory_usage = peak_memory = end_memory - start_memory
@@ -654,7 +743,7 @@ class SemaphoreDispatcher(BaseDispatcher):
         self,
         crawler: AsyncWebCrawler,  # noqa: F821
         urls: List[str],
-        config: CrawlerRunConfig,
+        config: Union[CrawlerRunConfig, List[CrawlerRunConfig]],
     ) -> List[CrawlerTaskResult]:
         self.crawler = crawler
         if self.monitor:

crawl4ai/async_logger.py

@@ -39,6 +39,7 @@ class LogColor(str, Enum):
     YELLOW = "yellow"
     MAGENTA = "magenta"
     DIM_MAGENTA = "dim magenta"
+    RED = "red"
 
     def __str__(self):
         """Automatically convert rich color to string."""

crawl4ai/async_url_seeder.py

@@ -424,10 +424,21 @@ class AsyncUrlSeeder:
         self._log("info", "Finished URL seeding for {domain}. Total URLs: {count}",
                   params={"domain": domain, "count": len(results)}, tag="URL_SEED")
 
-        # Sort by relevance score if query was provided
+        # Apply BM25 scoring if query was provided
         if query and extract_head and scoring_method == "bm25":
-            results.sort(key=lambda x: x.get(
-                "relevance_score", 0.0), reverse=True)
+            # Apply collective BM25 scoring across all documents
+            results = await self._apply_bm25_scoring(results, config)
+            
+            # Filter by score threshold if specified
+            if score_threshold is not None:
+                original_count = len(results)
+                results = [r for r in results if r.get("relevance_score", 0) >= score_threshold]
+                if original_count > len(results):
+                    self._log("info", "Filtered {filtered} URLs below score threshold {threshold}",
+                              params={"filtered": original_count - len(results), "threshold": score_threshold}, tag="URL_SEED")
+            
+            # Sort by relevance score
+            results.sort(key=lambda x: x.get("relevance_score", 0.0), reverse=True)
             self._log("info", "Sorted {count} URLs by relevance score for query: '{query}'",
                       params={"count": len(results), "query": query}, tag="URL_SEED")
         elif query and not extract_head:
@@ -818,7 +829,7 @@ class AsyncUrlSeeder:
 
     async def _iter_sitemap(self, url: str):
         try:
-            r = await self.client.get(url, timeout=15)
+            r = await self.client.get(url, timeout=15, follow_redirects=True)
             r.raise_for_status()
         except httpx.HTTPStatusError as e:
             self._log("warning", "Failed to fetch sitemap {url}: HTTP {status_code}",
@@ -982,28 +993,6 @@ class AsyncUrlSeeder:
                 "head_data": head_data,
             }
 
-            # Apply BM25 scoring if query is provided and head data exists
-            if query and ok and scoring_method == "bm25" and head_data:
-                text_context = self._extract_text_context(head_data)
-                if text_context:
-                    # Calculate BM25 score for this single document
-                    # scores = self._calculate_bm25_score(query, [text_context])
-                    scores = await asyncio.to_thread(self._calculate_bm25_score, query, [text_context])
-                    relevance_score = scores[0] if scores else 0.0
-                    entry["relevance_score"] = float(relevance_score)
-                else:
-                    # No text context, use URL-based scoring as fallback
-                    relevance_score = self._calculate_url_relevance_score(
-                        query, entry["url"])
-                    entry["relevance_score"] = float(relevance_score)
-            elif query:
-                # Query provided but no head data - we reject this entry
-                self._log("debug", "No head data for {url}, using URL-based scoring",
-                          params={"url": url}, tag="URL_SEED")
-                return
-                # relevance_score = self._calculate_url_relevance_score(query, entry["url"])
-                # entry["relevance_score"] = float(relevance_score)
-
         elif live:
             self._log("debug", "Performing live check for {url}", params={
                       "url": url}, tag="URL_SEED")
@@ -1013,35 +1002,13 @@ class AsyncUrlSeeder:
                       params={"status": status.upper(), "url": url}, tag="URL_SEED")
             entry = {"url": url, "status": status, "head_data": {}}
 
-            # Apply URL-based scoring if query is provided
-            if query:
-                relevance_score = self._calculate_url_relevance_score(
-                    query, url)
-                entry["relevance_score"] = float(relevance_score)
-
         else:
             entry = {"url": url, "status": "unknown", "head_data": {}}
 
-            # Apply URL-based scoring if query is provided
-            if query:
-                relevance_score = self._calculate_url_relevance_score(
-                    query, url)
-                entry["relevance_score"] = float(relevance_score)
-
-        # Now decide whether to add the entry based on score threshold
-        if query and "relevance_score" in entry:
-            if score_threshold is None or entry["relevance_score"] >= score_threshold:
-                if live or extract:
-                    await self._cache_set(cache_kind, url, entry)
-                res_list.append(entry)
-            else:
-                self._log("debug", "URL {url} filtered out with score {score} < {threshold}",
-                          params={"url": url, "score": entry["relevance_score"], "threshold": score_threshold}, tag="URL_SEED")
-        else:
-            # No query or no scoring - add as usual
-            if live or extract:
-                await self._cache_set(cache_kind, url, entry)
-            res_list.append(entry)
+        # Add entry to results (scoring will be done later)
+        if live or extract:
+            await self._cache_set(cache_kind, url, entry)
+        res_list.append(entry)
 
     async def _head_ok(self, url: str, timeout: int) -> bool:
         try:
@@ -1436,8 +1403,19 @@ class AsyncUrlSeeder:
             scores = bm25.get_scores(query_tokens)
 
             # Normalize scores to 0-1 range
-            max_score = max(scores) if max(scores) > 0 else 1.0
-            normalized_scores = [score / max_score for score in scores]
+            # BM25 can return negative scores, so we need to handle the full range
+            if len(scores) == 0:
+                return []
+            
+            min_score = min(scores)
+            max_score = max(scores)
+            
+            # If all scores are the same, return 0.5 for all
+            if max_score == min_score:
+                return [0.5] * len(scores)
+            
+            # Normalize to 0-1 range using min-max normalization
+            normalized_scores = [(score - min_score) / (max_score - min_score) for score in scores]
 
             return normalized_scores
         except Exception as e:

crawl4ai/async_webcrawler.py

@@ -363,7 +363,7 @@ class AsyncWebCrawler:
                         pdf_data=pdf_data,
                         verbose=config.verbose,
                         is_raw_html=True if url.startswith("raw:") else False,
-                        redirected_url=async_response.redirected_url, 
+                        redirected_url=async_response.redirected_url,
                         **kwargs,
                     )
 
@@ -509,7 +509,7 @@ class AsyncWebCrawler:
             tables = media.pop("tables", []) if isinstance(media, dict) else []
             links = result.links.model_dump() if hasattr(result.links, 'model_dump') else result.links
             metadata = result.metadata
-            
+
         fit_html = preprocess_html_for_schema(html_content=html, text_threshold= 500, max_size= 300_000)
 
         ################################
@@ -591,11 +591,13 @@ class AsyncWebCrawler:
             # Choose content based on input_format
             content_format = config.extraction_strategy.input_format
             if content_format == "fit_markdown" and not markdown_result.fit_markdown:
-                self.logger.warning(
-                    message="Fit markdown requested but not available. Falling back to raw markdown.",
-                    tag="EXTRACT",
-                    params={"url": _url},
-                )
+
+                self.logger.url_status(
+                        url=_url,
+                        success=bool(html),
+                        timing=time.perf_counter() - t1,
+                        tag="EXTRACT",
+                    )
                 content_format = "markdown"
 
             content = {
@@ -619,11 +621,12 @@ class AsyncWebCrawler:
             )
 
             # Log extraction completion
-            self.logger.info(
-                message="Completed for {url:.50}... | Time: {timing}s",
-                tag="EXTRACT",
-                params={"url": _url, "timing": time.perf_counter() - t1},
-            )
+            self.logger.url_status(
+                        url=_url,
+                        success=bool(html),
+                        timing=time.perf_counter() - t1,
+                        tag="EXTRACT",
+                    )
 
         # Apply HTML formatting if requested
         if config.prettiify:
@@ -650,7 +653,7 @@ class AsyncWebCrawler:
     async def arun_many(
         self,
         urls: List[str],
-        config: Optional[CrawlerRunConfig] = None,
+        config: Optional[Union[CrawlerRunConfig, List[CrawlerRunConfig]]] = None,
         dispatcher: Optional[BaseDispatcher] = None,
         # Legacy parameters maintained for backwards compatibility
         # word_count_threshold=MIN_WORD_THRESHOLD,
@@ -671,7 +674,9 @@ class AsyncWebCrawler:
 
         Args:
         urls: List of URLs to crawl
-        config: Configuration object controlling crawl behavior for all URLs
+        config: Configuration object(s) controlling crawl behavior. Can be:
+            - Single CrawlerRunConfig: Used for all URLs
+            - List[CrawlerRunConfig]: Configs with url_matcher for URL-specific settings
         dispatcher: The dispatcher strategy instance to use. Defaults to MemoryAdaptiveDispatcher
         [other parameters maintained for backwards compatibility]
 
@@ -736,7 +741,11 @@ class AsyncWebCrawler:
                 or task_result.result
             )
 
-        stream = config.stream
+        # Handle stream setting - use first config's stream setting if config is a list
+        if isinstance(config, list):
+            stream = config[0].stream if config else False
+        else:
+            stream = config.stream
 
         if stream:
 

crawl4ai/browser_adapter.py

@@ -0,0 +1,293 @@
+# browser_adapter.py
+"""
+Browser adapter for Crawl4AI to support both Playwright and undetected browsers
+with minimal changes to existing codebase.
+"""
+
+from abc import ABC, abstractmethod
+from typing import List, Dict, Any, Optional, Callable
+import time
+import json
+
+# Import both, but use conditionally
+try:
+    from playwright.async_api import Page
+except ImportError:
+    Page = Any
+
+try:
+    from patchright.async_api import Page as UndetectedPage
+except ImportError:
+    UndetectedPage = Any
+
+
+class BrowserAdapter(ABC):
+    """Abstract adapter for browser-specific operations"""
+    
+    @abstractmethod
+    async def evaluate(self, page: Page, expression: str, arg: Any = None) -> Any:
+        """Execute JavaScript in the page"""
+        pass
+    
+    @abstractmethod
+    async def setup_console_capture(self, page: Page, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup console message capturing, returns handler function if needed"""
+        pass
+    
+    @abstractmethod
+    async def setup_error_capture(self, page: Page, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup error capturing, returns handler function if needed"""
+        pass
+    
+    @abstractmethod
+    async def retrieve_console_messages(self, page: Page) -> List[Dict]:
+        """Retrieve captured console messages (for undetected browsers)"""
+        pass
+    
+    @abstractmethod
+    async def cleanup_console_capture(self, page: Page, handle_console: Optional[Callable], handle_error: Optional[Callable]):
+        """Clean up console event listeners"""
+        pass
+    
+    @abstractmethod
+    def get_imports(self) -> tuple:
+        """Get the appropriate imports for this adapter"""
+        pass
+
+
+class PlaywrightAdapter(BrowserAdapter):
+    """Adapter for standard Playwright"""
+    
+    async def evaluate(self, page: Page, expression: str, arg: Any = None) -> Any:
+        """Standard Playwright evaluate"""
+        if arg is not None:
+            return await page.evaluate(expression, arg)
+        return await page.evaluate(expression)
+    
+    async def setup_console_capture(self, page: Page, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup console capture using Playwright's event system"""
+        def handle_console_capture(msg):
+            try:
+                message_type = "unknown"
+                try:
+                    message_type = msg.type
+                except:
+                    pass
+                    
+                message_text = "unknown"
+                try:
+                    message_text = msg.text
+                except:
+                    pass
+                    
+                entry = {
+                    "type": message_type,
+                    "text": message_text,
+                    "timestamp": time.time()
+                }
+                
+                captured_console.append(entry)
+                
+            except Exception as e:
+                captured_console.append({
+                    "type": "console_capture_error", 
+                    "error": str(e), 
+                    "timestamp": time.time()
+                })
+        
+        page.on("console", handle_console_capture)
+        return handle_console_capture
+    
+    async def setup_error_capture(self, page: Page, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup error capture using Playwright's event system"""
+        def handle_pageerror_capture(err):
+            try:
+                error_message = "Unknown error"
+                try:
+                    error_message = err.message
+                except:
+                    pass
+                    
+                error_stack = ""
+                try:
+                    error_stack = err.stack
+                except:
+                    pass
+                    
+                captured_console.append({
+                    "type": "error",
+                    "text": error_message,
+                    "stack": error_stack,
+                    "timestamp": time.time()
+                })
+            except Exception as e:
+                captured_console.append({
+                    "type": "pageerror_capture_error", 
+                    "error": str(e), 
+                    "timestamp": time.time()
+                })
+        
+        page.on("pageerror", handle_pageerror_capture)
+        return handle_pageerror_capture
+    
+    async def retrieve_console_messages(self, page: Page) -> List[Dict]:
+        """Not needed for Playwright - messages are captured via events"""
+        return []
+    
+    async def cleanup_console_capture(self, page: Page, handle_console: Optional[Callable], handle_error: Optional[Callable]):
+        """Remove event listeners"""
+        if handle_console:
+            page.remove_listener("console", handle_console)
+        if handle_error:
+            page.remove_listener("pageerror", handle_error)
+    
+    def get_imports(self) -> tuple:
+        """Return Playwright imports"""
+        from playwright.async_api import Page, Error
+        from playwright.async_api import TimeoutError as PlaywrightTimeoutError
+        return Page, Error, PlaywrightTimeoutError
+
+
+class UndetectedAdapter(BrowserAdapter):
+    """Adapter for undetected browser automation with stealth features"""
+    
+    def __init__(self):
+        self._console_script_injected = {}
+    
+    async def evaluate(self, page: UndetectedPage, expression: str, arg: Any = None) -> Any:
+        """Undetected browser evaluate with isolated context"""
+        # For most evaluations, use isolated context for stealth
+        # Only use non-isolated when we need to access our injected console capture
+        isolated = not (
+            "__console" in expression or 
+            "__captured" in expression or
+            "__error" in expression or
+            "window.__" in expression
+        )
+        
+        if arg is not None:
+            return await page.evaluate(expression, arg, isolated_context=isolated)
+        return await page.evaluate(expression, isolated_context=isolated)
+    
+    async def setup_console_capture(self, page: UndetectedPage, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup console capture using JavaScript injection for undetected browsers"""
+        if not self._console_script_injected.get(page, False):
+            await page.add_init_script("""
+                // Initialize console capture
+                window.__capturedConsole = [];
+                window.__capturedErrors = [];
+                
+                // Store original console methods
+                const originalConsole = {};
+                ['log', 'info', 'warn', 'error', 'debug'].forEach(method => {
+                    originalConsole[method] = console[method];
+                    console[method] = function(...args) {
+                        try {
+                            window.__capturedConsole.push({
+                                type: method,
+                                text: args.map(arg => {
+                                    try {
+                                        if (typeof arg === 'object') {
+                                            return JSON.stringify(arg);
+                                        }
+                                        return String(arg);
+                                    } catch (e) {
+                                        return '[Object]';
+                                    }
+                                }).join(' '),
+                                timestamp: Date.now()
+                            });
+                        } catch (e) {
+                            // Fail silently to avoid detection
+                        }
+                        
+                        // Call original method
+                        originalConsole[method].apply(console, args);
+                    };
+                });
+            """)
+            self._console_script_injected[page] = True
+        
+        return None  # No handler function needed for undetected browser
+    
+    async def setup_error_capture(self, page: UndetectedPage, captured_console: List[Dict]) -> Optional[Callable]:
+        """Setup error capture using JavaScript injection for undetected browsers"""
+        if not self._console_script_injected.get(page, False):
+            await page.add_init_script("""
+                // Capture errors
+                window.addEventListener('error', (event) => {
+                    try {
+                        window.__capturedErrors.push({
+                            type: 'error',
+                            text: event.message,
+                            stack: event.error ? event.error.stack : '',
+                            filename: event.filename,
+                            lineno: event.lineno,
+                            colno: event.colno,
+                            timestamp: Date.now()
+                        });
+                    } catch (e) {
+                        // Fail silently
+                    }
+                });
+                
+                // Capture unhandled promise rejections
+                window.addEventListener('unhandledrejection', (event) => {
+                    try {
+                        window.__capturedErrors.push({
+                            type: 'unhandledrejection',
+                            text: event.reason ? String(event.reason) : 'Unhandled Promise Rejection',
+                            stack: event.reason && event.reason.stack ? event.reason.stack : '',
+                            timestamp: Date.now()
+                        });
+                    } catch (e) {
+                        // Fail silently
+                    }
+                });
+            """)
+            self._console_script_injected[page] = True
+        
+        return None  # No handler function needed for undetected browser
+    
+    async def retrieve_console_messages(self, page: UndetectedPage) -> List[Dict]:
+        """Retrieve captured console messages and errors from the page"""
+        messages = []
+        
+        try:
+            # Get console messages
+            console_messages = await page.evaluate(
+                "() => { const msgs = window.__capturedConsole || []; window.__capturedConsole = []; return msgs; }",
+                isolated_context=False
+            )
+            messages.extend(console_messages)
+            
+            # Get errors
+            errors = await page.evaluate(
+                "() => { const errs = window.__capturedErrors || []; window.__capturedErrors = []; return errs; }",
+                isolated_context=False
+            )
+            messages.extend(errors)
+            
+            # Convert timestamps from JS to Python format
+            for msg in messages:
+                if 'timestamp' in msg and isinstance(msg['timestamp'], (int, float)):
+                    msg['timestamp'] = msg['timestamp'] / 1000.0  # Convert from ms to seconds
+                    
+        except Exception:
+            # If retrieval fails, return empty list
+            pass
+        
+        return messages
+    
+    async def cleanup_console_capture(self, page: UndetectedPage, handle_console: Optional[Callable], handle_error: Optional[Callable]):
+        """Clean up for undetected browser - retrieve final messages"""
+        # For undetected browser, we don't have event listeners to remove
+        # but we should retrieve any final messages
+        final_messages = await self.retrieve_console_messages(page)
+        return final_messages
+    
+    def get_imports(self) -> tuple:
+        """Return undetected browser imports"""
+        from patchright.async_api import Page, Error
+        from patchright.async_api import TimeoutError as PlaywrightTimeoutError
+        return Page, Error, PlaywrightTimeoutError

crawl4ai/browser_manager.py

@@ -14,23 +14,8 @@ import hashlib
 from .js_snippet import load_js_script
 from .config import DOWNLOAD_PAGE_TIMEOUT
 from .async_configs import BrowserConfig, CrawlerRunConfig
-from playwright_stealth import StealthConfig
 from .utils import get_chromium_path
 
-stealth_config = StealthConfig(
-    webdriver=True,
-    chrome_app=True,
-    chrome_csi=True,
-    chrome_load_times=True,
-    chrome_runtime=True,
-    navigator_languages=True,
-    navigator_plugins=True,
-    navigator_permissions=True,
-    webgl_vendor=True,
-    outerdimensions=True,
-    navigator_hardware_concurrency=True,
-    media_codecs=True,
-)
 
 BROWSER_DISABLE_OPTIONS = [
     "--disable-background-networking",
@@ -588,21 +573,26 @@ class BrowserManager:
     _playwright_instance = None
     
     @classmethod
-    async def get_playwright(cls):
-        from playwright.async_api import async_playwright
+    async def get_playwright(cls, use_undetected: bool = False):
+        if use_undetected:
+            from patchright.async_api import async_playwright
+        else:
+            from playwright.async_api import async_playwright
         cls._playwright_instance = await async_playwright().start()
         return cls._playwright_instance    
 
-    def __init__(self, browser_config: BrowserConfig, logger=None):
+    def __init__(self, browser_config: BrowserConfig, logger=None, use_undetected: bool = False):
         """
         Initialize the BrowserManager with a browser configuration.
 
         Args:
             browser_config (BrowserConfig): Configuration object containing all browser settings
             logger: Logger instance for recording events and errors
+            use_undetected (bool): Whether to use undetected browser (Patchright)
         """
         self.config: BrowserConfig = browser_config
         self.logger = logger
+        self.use_undetected = use_undetected
 
         # Browser state
         self.browser = None
@@ -616,7 +606,11 @@ class BrowserManager:
 
         # Keep track of contexts by a "config signature," so each unique config reuses a single context
         self.contexts_by_config = {}
-        self._contexts_lock = asyncio.Lock() 
+        self._contexts_lock = asyncio.Lock()
+        
+        # Stealth-related attributes
+        self._stealth_instance = None
+        self._stealth_cm = None 
 
         # Initialize ManagedBrowser if needed
         if self.config.use_managed_browser:
@@ -645,9 +639,21 @@ class BrowserManager:
         if self.playwright is not None:
             await self.close()
             
-        from playwright.async_api import async_playwright
+        if self.use_undetected:
+            from patchright.async_api import async_playwright
+        else:
+            from playwright.async_api import async_playwright
 
-        self.playwright = await async_playwright().start()
+        # Initialize playwright with or without stealth
+        if self.config.enable_stealth and not self.use_undetected:
+            # Import stealth only when needed
+            from playwright_stealth import Stealth
+            # Use the recommended stealth wrapper approach
+            self._stealth_instance = Stealth()
+            self._stealth_cm = self._stealth_instance.use_async(async_playwright())
+            self.playwright = await self._stealth_cm.__aenter__()
+        else:
+            self.playwright = await async_playwright().start()
 
         if self.config.cdp_url or self.config.use_managed_browser:
             self.config.use_managed_browser = True
@@ -1109,5 +1115,19 @@ class BrowserManager:
             self.managed_browser = None
 
         if self.playwright:
-            await self.playwright.stop()
+            # Handle stealth context manager cleanup if it exists
+            if hasattr(self, '_stealth_cm') and self._stealth_cm is not None:
+                try:
+                    await self._stealth_cm.__aexit__(None, None, None)
+                except Exception as e:
+                    if self.logger:
+                        self.logger.error(
+                            message="Error closing stealth context: {error}",
+                            tag="ERROR", 
+                            params={"error": str(e)}
+                        )
+                self._stealth_cm = None
+                self._stealth_instance = None
+            else:
+                await self.playwright.stop()
             self.playwright = None

crawl4ai/browser_profiler.py

@@ -480,7 +480,7 @@ class BrowserProfiler:
                 self.logger.info("4. Exit", tag="MENU", base_color=LogColor.MAGENTA)
                 exit_option = "4"
             
-            self.logger.print(f"\n[cyan]Enter your choice (1-{exit_option}): [/cyan]", end="")
+            self.logger.info(f"\n[cyan]Enter your choice (1-{exit_option}): [/cyan]", end="")
             choice = input()
             
             if choice == "1":
@@ -637,9 +637,18 @@ class BrowserProfiler:
         self.logger.info(f"Debugging port: {debugging_port}", tag="CDP")
         self.logger.info(f"Headless mode: {headless}", tag="CDP")
         
+        # create browser config
+        browser_config = BrowserConfig(
+            browser_type=browser_type,
+            headless=headless,
+            user_data_dir=profile_path,
+            debugging_port=debugging_port,
+            verbose=True
+        )
+        
         # Create managed browser instance
         managed_browser = ManagedBrowser(
-            browser_type=browser_type,
+            browser_config=browser_config,
             user_data_dir=profile_path,
             headless=headless,
             logger=self.logger,

crawl4ai/content_scraping_strategy.py

@@ -98,20 +98,20 @@ class ContentScrapingStrategy(ABC):
         pass
 
 
-class WebScrapingStrategy(ContentScrapingStrategy):
+class LXMLWebScrapingStrategy(ContentScrapingStrategy):
     """
-    Class for web content scraping. Perhaps the most important class.
-
-    How it works:
-    1. Extract content from HTML using BeautifulSoup.
-    2. Clean the extracted content using a content cleaning strategy.
-    3. Filter the cleaned content using a content filtering strategy.
-    4. Generate markdown content from the filtered content.
-    5. Return the markdown content.
+    LXML-based implementation for fast web content scraping.
+    
+    This is the primary scraping strategy in Crawl4AI, providing high-performance
+    HTML parsing and content extraction using the lxml library.
+    
+    Note: WebScrapingStrategy is now an alias for this class to maintain
+    backward compatibility.
     """
-
     def __init__(self, logger=None):
         self.logger = logger
+        self.DIMENSION_REGEX = re.compile(r"(\d+)(\D*)")
+        self.BASE64_PATTERN = re.compile(r'data:image/[^;]+;base64,([^"]+)')
 
     def _log(self, level, message, tag="SCRAPE", **kwargs):
         """Helper method to safely use logger."""
@@ -132,7 +132,7 @@ class WebScrapingStrategy(ContentScrapingStrategy):
             ScrapingResult: A structured result containing the scraped content.
         """
         actual_url = kwargs.get("redirected_url", url)
-        raw_result = self._scrap(actual_url, html, is_async=False, **kwargs)
+        raw_result = self._scrap(actual_url, html, **kwargs)
         if raw_result is None:
             return ScrapingResult(
                 cleaned_html="",
@@ -196,376 +196,9 @@ class WebScrapingStrategy(ContentScrapingStrategy):
         Returns:
             ScrapingResult: A structured result containing the scraped content.
         """
-        return await asyncio.to_thread(self._scrap, url, html, **kwargs)
+        return await asyncio.to_thread(self.scrap, url, html, **kwargs)
 
-    def is_data_table(self, table: Tag, **kwargs) -> bool:
-        """
-        Determine if a table element is a data table (not a layout table).
-
-        Args:
-            table (Tag): BeautifulSoup Tag representing a table element
-            **kwargs: Additional keyword arguments including table_score_threshold
-
-        Returns:
-            bool: True if the table is a data table, False otherwise
-        """
-        score = 0
-        
-        # Check for thead and tbody
-        has_thead = len(table.select('thead')) > 0
-        has_tbody = len(table.select('tbody')) > 0
-        if has_thead:
-            score += 2
-        if has_tbody:
-            score += 1
-            
-        # Check for th elements
-        th_count = len(table.select('th'))
-        if th_count > 0:
-            score += 2
-            if has_thead or len(table.select('tr:first-child th')) > 0:
-                score += 1
-                
-        # Check for nested tables
-        if len(table.select('table')) > 0:
-            score -= 3
-            
-        # Role attribute check
-        role = table.get('role', '').lower()
-        if role in {'presentation', 'none'}:
-            score -= 3
-            
-        # Column consistency
-        rows = table.select('tr')
-        if not rows:
-            return False
-            
-        col_counts = [len(row.select('td, th')) for row in rows]
-        avg_cols = sum(col_counts) / len(col_counts)
-        variance = sum((c - avg_cols)**2 for c in col_counts) / len(col_counts)
-        if variance < 1:
-            score += 2
-            
-        # Caption and summary
-        if table.select('caption'):
-            score += 2
-        if table.has_attr('summary') and table['summary']:
-            score += 1
-            
-        # Text density
-        total_text = sum(len(cell.get_text().strip()) for row in rows for cell in row.select('td, th'))
-        total_tags = sum(1 for _ in table.descendants if isinstance(_, Tag))
-        text_ratio = total_text / (total_tags + 1e-5)
-        if text_ratio > 20:
-            score += 3
-        elif text_ratio > 10:
-            score += 2
-            
-        # Data attributes
-        data_attrs = sum(1 for attr in table.attrs if attr.startswith('data-'))
-        score += data_attrs * 0.5
-        
-        # Size check
-        if avg_cols >= 2 and len(rows) >= 2:
-            score += 2
-            
-        threshold = kwargs.get('table_score_threshold', 7)
-        return score >= threshold
-    
-    def extract_table_data(self, table: Tag) -> dict:
-        """
-        Extract structured data from a table element.
-        
-        Args:
-            table (Tag): BeautifulSoup Tag representing a table element
-            
-        Returns:
-            dict: Dictionary containing table data (headers, rows, caption, summary)
-        """
-        caption_elem = table.select_one('caption')
-        caption = caption_elem.get_text().strip() if caption_elem else ""
-        summary = table.get('summary', '').strip()
-        
-        # Extract headers with colspan handling
-        headers = []
-        thead_rows = table.select('thead tr')
-        if thead_rows:
-            header_cells = thead_rows[0].select('th')
-            for cell in header_cells:
-                text = cell.get_text().strip()
-                colspan = int(cell.get('colspan', 1))
-                headers.extend([text] * colspan)
-        else:
-            first_row = table.select('tr:first-child')
-            if first_row:
-                for cell in first_row[0].select('th, td'):
-                    text = cell.get_text().strip()
-                    colspan = int(cell.get('colspan', 1))
-                    headers.extend([text] * colspan)
-        
-        # Extract rows with colspan handling
-        rows = []
-        all_rows = table.select('tr')
-        thead = table.select_one('thead')
-        tbody_rows = []
-
-        if thead:
-            thead_rows = thead.select('tr')
-            tbody_rows = [row for row in all_rows if row not in thead_rows]
-        else:
-            if all_rows and all_rows[0].select('th'):
-                tbody_rows = all_rows[1:]
-            else:
-                tbody_rows = all_rows
-                
-        for row in tbody_rows:        
-        # for row in table.select('tr:not(:has(ancestor::thead))'):
-            row_data = []
-            for cell in row.select('td'):
-                text = cell.get_text().strip()
-                colspan = int(cell.get('colspan', 1))
-                row_data.extend([text] * colspan)
-            if row_data:
-                rows.append(row_data)
-                
-        # Align rows with headers
-        max_columns = len(headers) if headers else (max(len(row) for row in rows) if rows else 0)
-        aligned_rows = []
-        for row in rows:
-            aligned = row[:max_columns] + [''] * (max_columns - len(row))
-            aligned_rows.append(aligned)
-            
-        if not headers:
-            headers = [f"Column {i+1}" for i in range(max_columns)]
-            
-        return {
-            "headers": headers,
-            "rows": aligned_rows,
-            "caption": caption,
-            "summary": summary,
-        }
-    
-    def flatten_nested_elements(self, node):
-        """
-        Flatten nested elements in a HTML tree.
-
-        Args:
-            node (Tag): The root node of the HTML tree.
-
-        Returns:
-            Tag: The flattened HTML tree.
-        """
-        if isinstance(node, NavigableString):
-            return node
-        if (
-            len(node.contents) == 1
-            and isinstance(node.contents[0], Tag)
-            and node.contents[0].name == node.name
-        ):
-            return self.flatten_nested_elements(node.contents[0])
-        node.contents = [self.flatten_nested_elements(child) for child in node.contents]
-        return node
-
-    def find_closest_parent_with_useful_text(self, tag, **kwargs):
-        """
-        Find the closest parent with useful text.
-
-        Args:
-            tag (Tag): The starting tag to search from.
-            **kwargs: Additional keyword arguments.
-
-        Returns:
-            Tag: The closest parent with useful text, or None if not found.
-        """
-        image_description_min_word_threshold = kwargs.get(
-            "image_description_min_word_threshold", IMAGE_DESCRIPTION_MIN_WORD_THRESHOLD
-        )
-        current_tag = tag
-        while current_tag:
-            current_tag = current_tag.parent
-            # Get the text content of the parent tag
-            if current_tag:
-                text_content = current_tag.get_text(separator=" ", strip=True)
-                # Check if the text content has at least word_count_threshold
-                if len(text_content.split()) >= image_description_min_word_threshold:
-                    return text_content
-        return None
-
-    def remove_unwanted_attributes(
-        self, element, important_attrs, keep_data_attributes=False
-    ):
-        """
-        Remove unwanted attributes from an HTML element.
-
-        Args:
-            element (Tag): The HTML element to remove attributes from.
-            important_attrs (list): List of important attributes to keep.
-            keep_data_attributes (bool): Whether to keep data attributes.
-
-        Returns:
-            None
-        """
-        attrs_to_remove = []
-        for attr in element.attrs:
-            if attr not in important_attrs:
-                if keep_data_attributes:
-                    if not attr.startswith("data-"):
-                        attrs_to_remove.append(attr)
-                else:
-                    attrs_to_remove.append(attr)
-
-        for attr in attrs_to_remove:
-            del element[attr]
-
-    def process_image(self, img, url, index, total_images, **kwargs):
-        """
-        Process an image element.
-
-        How it works:
-        1. Check if the image has valid display and inside undesired html elements.
-        2. Score an image for it's usefulness.
-        3. Extract image file metadata to extract size and extension.
-        4. Generate a dictionary with the processed image information.
-        5. Return the processed image information.
-
-        Args:
-            img (Tag): The image element to process.
-            url (str): The URL of the page containing the image.
-            index (int): The index of the image in the list of images.
-            total_images (int): The total number of images in the list.
-            **kwargs: Additional keyword arguments.
-
-        Returns:
-            dict: A dictionary containing the processed image information.
-        """
-        # parse_srcset = lambda s: [{'url': u.strip().split()[0], 'width': u.strip().split()[-1].rstrip('w')
-        #                 if ' ' in u else None}
-        #                 for u in [f"http{p}" for p in s.split("http") if p]]
-
-        # Constants for checks
-        classes_to_check = frozenset(["button", "icon", "logo"])
-        tags_to_check = frozenset(["button", "input"])
-        image_formats = frozenset(["jpg", "jpeg", "png", "webp", "avif", "gif"])
-
-        # Pre-fetch commonly used attributes
-        style = img.get("style", "")
-        alt = img.get("alt", "")
-        src = img.get("src", "")
-        data_src = img.get("data-src", "")
-        srcset = img.get("srcset", "")
-        data_srcset = img.get("data-srcset", "")
-        width = img.get("width")
-        height = img.get("height")
-        parent = img.parent
-        parent_classes = parent.get("class", [])
-
-        # Quick validation checks
-        if (
-            "display:none" in style
-            or parent.name in tags_to_check
-            or any(c in cls for c in parent_classes for cls in classes_to_check)
-            or any(c in src for c in classes_to_check)
-            or any(c in alt for c in classes_to_check)
-        ):
-            return None
-
-        # Quick score calculation
-        score = 0
-        if width and width.isdigit():
-            width_val = int(width)
-            score += 1 if width_val > 150 else 0
-        if height and height.isdigit():
-            height_val = int(height)
-            score += 1 if height_val > 150 else 0
-        if alt:
-            score += 1
-        score += index / total_images < 0.5
-
-        # image_format = ''
-        # if "data:image/" in src:
-        #     image_format = src.split(',')[0].split(';')[0].split('/')[1].split(';')[0]
-        # else:
-        #     image_format = os.path.splitext(src)[1].lower().strip('.').split('?')[0]
-
-        # if image_format in ('jpg', 'png', 'webp', 'avif'):
-        #     score += 1
-
-        # Check for image format in all possible sources
-        def has_image_format(url):
-            return any(fmt in url.lower() for fmt in image_formats)
-
-        # Score for having proper image sources
-        if any(has_image_format(url) for url in [src, data_src, srcset, data_srcset]):
-            score += 1
-        if srcset or data_srcset:
-            score += 1
-        if img.find_parent("picture"):
-            score += 1
-
-        # Detect format from any available source
-        detected_format = None
-        for url in [src, data_src, srcset, data_srcset]:
-            if url:
-                format_matches = [fmt for fmt in image_formats if fmt in url.lower()]
-                if format_matches:
-                    detected_format = format_matches[0]
-                    break
-
-        if score <= kwargs.get("image_score_threshold", IMAGE_SCORE_THRESHOLD):
-            return None
-
-        # Use set for deduplication
-        unique_urls = set()
-        image_variants = []
-
-        # Generate a unique group ID for this set of variants
-        group_id = index
-
-        # Base image info template
-        base_info = {
-            "alt": alt,
-            "desc": self.find_closest_parent_with_useful_text(img, **kwargs),
-            "score": score,
-            "type": "image",
-            "group_id": group_id,  # Group ID for this set of variants
-            "format": detected_format,
-        }
-
-        # Inline function for adding variants
-        def add_variant(src, width=None):
-            if src and not src.startswith("data:") and src not in unique_urls:
-                unique_urls.add(src)
-                image_variants.append({**base_info, "src": src, "width": width})
-
-        # Process all sources
-        add_variant(src)
-        add_variant(data_src)
-
-        # Handle srcset and data-srcset in one pass
-        for attr in ("srcset", "data-srcset"):
-            if value := img.get(attr):
-                for source in parse_srcset(value):
-                    add_variant(source["url"], source["width"])
-
-        # Quick picture element check
-        if picture := img.find_parent("picture"):
-            for source in picture.find_all("source"):
-                if srcset := source.get("srcset"):
-                    for src in parse_srcset(srcset):
-                        add_variant(src["url"], src["width"])
-
-        # Framework-specific attributes in one pass
-        for attr, value in img.attrs.items():
-            if (
-                attr.startswith("data-")
-                and ("src" in attr or "srcset" in attr)
-                and "http" in value
-            ):
-                add_variant(value)
-
-        return image_variants if image_variants else None
-
-    def process_element(self, url, element: PageElement, **kwargs) -> Dict[str, Any]:
+    def process_element(self, url, element: lhtml.HtmlElement, **kwargs) -> Dict[str, Any]:
         """
         Process an HTML element.
 
@@ -577,7 +210,7 @@ class WebScrapingStrategy(ContentScrapingStrategy):
 
         Args:
             url (str): The URL of the page containing the element.
-            element (Tag): The HTML element to process.
+            element (lhtml.HtmlElement): The HTML element to process.
             **kwargs: Additional keyword arguments.
 
         Returns:
@@ -595,509 +228,6 @@ class WebScrapingStrategy(ContentScrapingStrategy):
             "external_links_dict": external_links_dict,
         }
 
-    def _process_element(
-        self,
-        url,
-        element: PageElement,
-        media: Dict[str, Any],
-        internal_links_dict: Dict[str, Any],
-        external_links_dict: Dict[str, Any],
-        **kwargs,
-    ) -> bool:
-        """
-        Process an HTML element.
-        """
-        try:
-            if isinstance(element, NavigableString):
-                if isinstance(element, Comment):
-                    element.extract()
-                return False
-
-            # if element.name == 'img':
-            #     process_image(element, url, 0, 1)
-            #     return True
-            base_domain = kwargs.get("base_domain", get_base_domain(url))
-
-            if element.name in ["script", "style", "link", "meta", "noscript"]:
-                element.decompose()
-                return False
-
-            keep_element = False
-            # Special case for table elements - always preserve structure
-            if element.name in ["tr", "td", "th"]:
-                keep_element = True
-
-            exclude_domains = kwargs.get("exclude_domains", [])
-            # exclude_social_media_domains = kwargs.get('exclude_social_media_domains', set(SOCIAL_MEDIA_DOMAINS))
-            # exclude_social_media_domains = SOCIAL_MEDIA_DOMAINS + kwargs.get('exclude_social_media_domains', [])
-            # exclude_social_media_domains = list(set(exclude_social_media_domains))
-
-            try:
-                if element.name == "a" and element.get("href"):
-                    href = element.get("href", "").strip()
-                    if not href:  # Skip empty hrefs
-                        return False
-
-                    # url_base = url.split("/")[2]
-
-                    # Normalize the URL
-                    try:
-                        normalized_href = normalize_url(href, url)
-                    except ValueError:
-                        # logging.warning(f"Invalid URL format: {href}, Error: {str(e)}")
-                        return False
-
-                    link_data = {
-                        "href": normalized_href,
-                        "text": element.get_text().strip(),
-                        "title": element.get("title", "").strip(),
-                        "base_domain": base_domain,
-                    }
-
-                    is_external = is_external_url(normalized_href, base_domain)
-
-                    keep_element = True
-
-                    # Handle external link exclusions
-                    if is_external:
-                        link_base_domain = get_base_domain(normalized_href)
-                        link_data["base_domain"] = link_base_domain
-                        if kwargs.get("exclude_external_links", False):
-                            element.decompose()
-                            return False
-                        # elif kwargs.get('exclude_social_media_links', False):
-                        #     if link_base_domain in exclude_social_media_domains:
-                        #         element.decompose()
-                        #         return False
-                        # if any(domain in normalized_href.lower() for domain in exclude_social_media_domains):
-                        #     element.decompose()
-                        #     return False
-                        elif exclude_domains:
-                            if link_base_domain in exclude_domains:
-                                element.decompose()
-                                return False
-                            # if any(domain in normalized_href.lower() for domain in kwargs.get('exclude_domains', [])):
-                            #     element.decompose()
-                            #     return False
-
-                    if is_external:
-                        if normalized_href not in external_links_dict:
-                            external_links_dict[normalized_href] = link_data
-                    else:
-                        if kwargs.get("exclude_internal_links", False):
-                            element.decompose()
-                            return False
-                        if normalized_href not in internal_links_dict:
-                            internal_links_dict[normalized_href] = link_data
-
-            except Exception as e:
-                raise Exception(f"Error processing links: {str(e)}")
-
-            try:
-                if element.name == "img":
-                    potential_sources = [
-                        "src",
-                        "data-src",
-                        "srcset" "data-lazy-src",
-                        "data-original",
-                    ]
-                    src = element.get("src", "")
-                    while not src and potential_sources:
-                        src = element.get(potential_sources.pop(0), "")
-                    if not src:
-                        element.decompose()
-                        return False
-
-                    # If it is srcset pick up the first image
-                    if "srcset" in element.attrs:
-                        src = element.attrs["srcset"].split(",")[0].split(" ")[0]
-
-                    # If image src is internal, then skip
-                    if not is_external_url(src, base_domain):
-                        return True
-
-                    image_src_base_domain = get_base_domain(src)
-
-                    # Check flag if we should remove external images
-                    if kwargs.get("exclude_external_images", False):
-                        element.decompose()
-                        return False
-                        # src_url_base = src.split('/')[2]
-                        # url_base = url.split('/')[2]
-                        # if url_base not in src_url_base:
-                        #     element.decompose()
-                        #     return False
-
-                    # if kwargs.get('exclude_social_media_links', False):
-                    #     if image_src_base_domain in exclude_social_media_domains:
-                    #         element.decompose()
-                    #         return False
-                    # src_url_base = src.split('/')[2]
-                    # url_base = url.split('/')[2]
-                    # if any(domain in src for domain in exclude_social_media_domains):
-                    #     element.decompose()
-                    #     return False
-
-                    # Handle exclude domains
-                    if exclude_domains:
-                        if image_src_base_domain in exclude_domains:
-                            element.decompose()
-                            return False
-                        # if any(domain in src for domain in kwargs.get('exclude_domains', [])):
-                        #     element.decompose()
-                        #     return False
-
-                    return True  # Always keep image elements
-            except Exception:
-                raise "Error processing images"
-
-            # Check if flag to remove all forms is set
-            if kwargs.get("remove_forms", False) and element.name == "form":
-                element.decompose()
-                return False
-
-            if element.name in ["video", "audio"]:
-                media[f"{element.name}s"].append(
-                    {
-                        "src": element.get("src"),
-                        "alt": element.get("alt"),
-                        "type": element.name,
-                        "description": self.find_closest_parent_with_useful_text(
-                            element, **kwargs
-                        ),
-                    }
-                )
-                source_tags = element.find_all("source")
-                for source_tag in source_tags:
-                    media[f"{element.name}s"].append(
-                        {
-                            "src": source_tag.get("src"),
-                            "alt": element.get("alt"),
-                            "type": element.name,
-                            "description": self.find_closest_parent_with_useful_text(
-                                element, **kwargs
-                            ),
-                        }
-                    )
-                return True  # Always keep video and audio elements
-
-            if element.name in ONLY_TEXT_ELIGIBLE_TAGS:
-                if kwargs.get("only_text", False):
-                    element.replace_with(element.get_text())
-
-            try:
-                self.remove_unwanted_attributes(
-                    element, IMPORTANT_ATTRS + kwargs.get("keep_attrs", []) , kwargs.get("keep_data_attributes", False)
-                )
-            except Exception as e:
-                # print('Error removing unwanted attributes:', str(e))
-                self._log(
-                    "error",
-                    message="Error removing unwanted attributes: {error}",
-                    tag="SCRAPE",
-                    params={"error": str(e)},
-                )
-            # Process children
-            for child in list(element.children):
-                if isinstance(child, NavigableString) and not isinstance(
-                    child, Comment
-                ):
-                    if len(child.strip()) > 0:
-                        keep_element = True
-                else:
-                    if self._process_element(
-                        url,
-                        child,
-                        media,
-                        internal_links_dict,
-                        external_links_dict,
-                        **kwargs,
-                    ):
-                        keep_element = True
-
-            # Check word count
-            word_count_threshold = kwargs.get(
-                "word_count_threshold", MIN_WORD_THRESHOLD
-            )
-            if not keep_element:
-                word_count = len(element.get_text(strip=True).split())
-                keep_element = word_count >= word_count_threshold
-
-            if not keep_element:
-                element.decompose()
-
-            return keep_element
-        except Exception as e:
-            # print('Error processing element:', str(e))
-            self._log(
-                "error",
-                message="Error processing element: {error}",
-                tag="SCRAPE",
-                params={"error": str(e)},
-            )
-            return False
-
-    def _scrap(
-        self,
-        url: str,
-        html: str,
-        word_count_threshold: int = MIN_WORD_THRESHOLD,
-        css_selector: str = None,
-        target_elements: List[str] = None,
-        **kwargs,
-    ) -> Dict[str, Any]:
-        """
-        Extract content from HTML using BeautifulSoup.
-
-        Args:
-            url (str): The URL of the page to scrape.
-            html (str): The HTML content of the page to scrape.
-            word_count_threshold (int): The minimum word count threshold for content extraction.
-            css_selector (str): The CSS selector to use for content extraction.
-            **kwargs: Additional keyword arguments.
-
-        Returns:
-            dict: A dictionary containing the extracted content.
-        """
-        success = True
-        if not html:
-            return None
-
-        parser_type = kwargs.get("parser", "lxml")
-        soup = BeautifulSoup(html, parser_type)
-        body = soup.body
-        if body is None:
-            raise Exception("'<body>' tag is not found in fetched html. Consider adding wait_for=\"css:body\" to wait for body tag to be loaded into DOM.")
-        base_domain = get_base_domain(url)
-        
-        # Early removal of all images if exclude_all_images is set
-        # This happens before any processing to minimize memory usage
-        if kwargs.get("exclude_all_images", False):
-            for img in body.find_all('img'):
-                img.decompose()
-
-        try:
-            meta = extract_metadata("", soup)
-        except Exception as e:
-            self._log(
-                "error",
-                message="Error extracting metadata: {error}",
-                tag="SCRAPE",
-                params={"error": str(e)},
-            )
-            meta = {}
-
-        # Handle tag-based removal first - faster than CSS selection
-        excluded_tags = set(kwargs.get("excluded_tags", []) or [])
-        if excluded_tags:
-            for element in body.find_all(lambda tag: tag.name in excluded_tags):
-                element.extract()
-
-        # Handle CSS selector-based removal
-        excluded_selector = kwargs.get("excluded_selector", "")
-        if excluded_selector:
-            is_single_selector = (
-                "," not in excluded_selector and " " not in excluded_selector
-            )
-            if is_single_selector:
-                while element := body.select_one(excluded_selector):
-                    element.extract()
-            else:
-                for element in body.select(excluded_selector):
-                    element.extract()
-
-        content_element = None
-        if target_elements:
-            try:
-                for_content_targeted_element = []
-                for target_element in target_elements:
-                    for_content_targeted_element.extend(body.select(target_element))
-                content_element = soup.new_tag("div")
-                for el in for_content_targeted_element:
-                    content_element.append(copy.deepcopy(el))
-            except Exception as e:
-                self._log("error", f"Error with target element detection: {str(e)}", "SCRAPE")
-                return None
-        else:
-            content_element = body     
-
-        kwargs["exclude_social_media_domains"] = set(
-            kwargs.get("exclude_social_media_domains", []) + SOCIAL_MEDIA_DOMAINS
-        )
-        kwargs["exclude_domains"] = set(kwargs.get("exclude_domains", []))
-        if kwargs.get("exclude_social_media_links", False):
-            kwargs["exclude_domains"] = kwargs["exclude_domains"].union(
-                kwargs["exclude_social_media_domains"]
-            )
-
-        result_obj = self.process_element(
-            url,
-            body,
-            word_count_threshold=word_count_threshold,
-            base_domain=base_domain,
-            **kwargs,
-        )
-
-        links = {"internal": [], "external": []}
-        media = result_obj["media"]
-        internal_links_dict = result_obj["internal_links_dict"]
-        external_links_dict = result_obj["external_links_dict"]
-
-        # Update the links dictionary with unique links
-        links["internal"] = list(internal_links_dict.values())
-        links["external"] = list(external_links_dict.values())
-        
-        # Extract head content for links if configured
-        link_preview_config = kwargs.get("link_preview_config")
-        if link_preview_config is not None:
-            try:
-                import asyncio
-                from .link_preview import LinkPreview
-                from .models import Links, Link
-                
-                verbose = link_preview_config.verbose
-                
-                if verbose:
-                    self._log("info", "Starting link head extraction for {internal} internal and {external} external links",
-                              params={"internal": len(links["internal"]), "external": len(links["external"])}, tag="LINK_EXTRACT")
-                
-                # Convert dict links to Link objects
-                internal_links = [Link(**link_data) for link_data in links["internal"]]
-                external_links = [Link(**link_data) for link_data in links["external"]]
-                links_obj = Links(internal=internal_links, external=external_links)
-                
-                # Create a config object for LinkPreview  
-                class TempCrawlerRunConfig:
-                    def __init__(self, link_config, score_links):
-                        self.link_preview_config = link_config
-                        self.score_links = score_links
-                
-                config = TempCrawlerRunConfig(link_preview_config, kwargs.get("score_links", False))
-                
-                # Extract head content (run async operation in sync context)
-                async def extract_links():
-                    async with LinkPreview(self.logger) as extractor:
-                        return await extractor.extract_link_heads(links_obj, config)
-                
-                # Run the async operation
-                try:
-                    # Check if we're already in an async context
-                    loop = asyncio.get_running_loop()
-                    # If we're in an async context, we need to run in a thread
-                    import concurrent.futures
-                    with concurrent.futures.ThreadPoolExecutor() as executor:
-                        future = executor.submit(asyncio.run, extract_links())
-                        updated_links = future.result()
-                except RuntimeError:
-                    # No running loop, we can use asyncio.run directly
-                    updated_links = asyncio.run(extract_links())
-                
-                # Convert back to dict format
-                links["internal"] = [link.dict() for link in updated_links.internal]
-                links["external"] = [link.dict() for link in updated_links.external]
-                
-                if verbose:
-                    successful_internal = len([l for l in updated_links.internal if l.head_extraction_status == "valid"])
-                    successful_external = len([l for l in updated_links.external if l.head_extraction_status == "valid"])
-                    self._log("info", "Link head extraction completed: {internal_success}/{internal_total} internal, {external_success}/{external_total} external",
-                              params={
-                                  "internal_success": successful_internal,
-                                  "internal_total": len(updated_links.internal),
-                                  "external_success": successful_external,
-                                  "external_total": len(updated_links.external)
-                              }, tag="LINK_EXTRACT")
-                else:
-                    self._log("info", "Link head extraction completed successfully", tag="LINK_EXTRACT")
-                
-            except Exception as e:
-                self._log("error", f"Link head extraction failed: {str(e)}", tag="LINK_EXTRACT")
-                # Continue with original links if extraction fails
-
-        # # Process images using ThreadPoolExecutor
-        imgs = body.find_all("img")
-
-        media["images"] = [
-            img
-            for result in (
-                self.process_image(img, url, i, len(imgs), **kwargs)
-                for i, img in enumerate(imgs)
-            )
-            if result is not None
-            for img in result
-        ]
-        
-        # Process tables if not excluded
-        excluded_tags = set(kwargs.get("excluded_tags", []) or [])
-        if 'table' not in excluded_tags:
-            tables = body.find_all('table')
-            for table in tables:
-                if self.is_data_table(table, **kwargs):
-                    table_data = self.extract_table_data(table)
-                    media["tables"].append(table_data)
-
-        body = self.flatten_nested_elements(body)
-        base64_pattern = re.compile(r'data:image/[^;]+;base64,([^"]+)')
-        for img in imgs:
-            src = img.get("src", "")
-            if base64_pattern.match(src):
-                # Replace base64 data with empty string
-                img["src"] = base64_pattern.sub("", src)
-
-        str_body = ""
-        try:
-            str_body = content_element.encode_contents().decode("utf-8")
-        except Exception:
-            # Reset body to the original HTML
-            success = False
-            body = BeautifulSoup(html, "html.parser")
-
-            # Create a new div with a special ID
-            error_div = body.new_tag("div", id="crawl4ai_error_message")
-            error_div.string = """
-            Crawl4AI Error: This page is not fully supported.
-            
-            Possible reasons:
-            1. The page may have restrictions that prevent crawling.
-            2. The page might not be fully loaded.
-            
-            Suggestions:
-            - Try calling the crawl function with these parameters:
-            magic=True,
-            - Set headless=False to visualize what's happening on the page.
-            
-            If the issue persists, please check the page's structure and any potential anti-crawling measures.
-            """
-
-            # Append the error div to the body
-            body.append(error_div)
-            str_body = body.encode_contents().decode("utf-8")
-
-            print(
-                "[LOG] 😧 Error: After processing the crawled HTML and removing irrelevant tags, nothing was left in the page. Check the markdown for further details."
-            )
-            self._log(
-                "error",
-                message="After processing the crawled HTML and removing irrelevant tags, nothing was left in the page. Check the markdown for further details.",
-                tag="SCRAPE",
-            )
-
-        cleaned_html = str_body.replace("\n\n", "\n").replace("  ", " ")
-
-        return {
-            "cleaned_html": cleaned_html,
-            "success": success,
-            "media": media,
-            "links": links,
-            "metadata": meta,
-        }
-
-
-class LXMLWebScrapingStrategy(WebScrapingStrategy):
-    def __init__(self, logger=None):
-        super().__init__(logger)
-        self.DIMENSION_REGEX = re.compile(r"(\d+)(\D*)")
-        self.BASE64_PATTERN = re.compile(r'data:image/[^;]+;base64,([^"]+)')
-
     def _process_element(
         self,
         url: str,
@@ -1140,10 +270,10 @@ class LXMLWebScrapingStrategy(WebScrapingStrategy):
                         link_data["intrinsic_score"] = intrinsic_score
                     except Exception:
                         # Fail gracefully - assign default score
-                        link_data["intrinsic_score"] = float('inf')
+                        link_data["intrinsic_score"] = 0
                 else:
                     # No scoring enabled - assign infinity (all links equal priority)
-                    link_data["intrinsic_score"] = float('inf')
+                    link_data["intrinsic_score"] = 0
 
                 is_external = is_external_url(normalized_href, base_domain)
                 if is_external:
@@ -1857,3 +987,7 @@ class LXMLWebScrapingStrategy(WebScrapingStrategy):
                 "links": {"internal": [], "external": []},
                 "metadata": {},
             }
+
+
+# Backward compatibility alias
+WebScrapingStrategy = LXMLWebScrapingStrategy

crawl4ai/deep_crawling/bff_strategy.py

@@ -116,11 +116,6 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
                 
             valid_links.append(base_url)
             
-        # If we have more valid links than capacity, limit them
-        if len(valid_links) > remaining_capacity:
-            valid_links = valid_links[:remaining_capacity]
-            self.logger.info(f"Limiting to {remaining_capacity} URLs due to max_pages limit")
-            
         # Record the new depths and add to next_links
         for url in valid_links:
             depths[url] = new_depth
@@ -140,7 +135,8 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
         """
         queue: asyncio.PriorityQueue = asyncio.PriorityQueue()
         # Push the initial URL with score 0 and depth 0.
-        await queue.put((0, 0, start_url, None))
+        initial_score = self.url_scorer.score(start_url) if self.url_scorer else 0
+        await queue.put((-initial_score, 0, start_url, None))
         visited: Set[str] = set()
         depths: Dict[str, int] = {start_url: 0}
 
@@ -150,6 +146,14 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
                 self.logger.info(f"Max pages limit ({self.max_pages}) reached, stopping crawl")
                 break
                 
+            # Calculate how many more URLs we can process in this batch
+            remaining = self.max_pages - self._pages_crawled
+            batch_size = min(BATCH_SIZE, remaining)
+            if batch_size <= 0:
+                # No more pages to crawl
+                self.logger.info(f"Max pages limit ({self.max_pages}) reached, stopping crawl")
+                break
+                
             batch: List[Tuple[float, int, str, Optional[str]]] = []
             # Retrieve up to BATCH_SIZE items from the priority queue.
             for _ in range(BATCH_SIZE):
@@ -179,11 +183,15 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
                 result.metadata = result.metadata or {}
                 result.metadata["depth"] = depth
                 result.metadata["parent_url"] = parent_url
-                result.metadata["score"] = score
+                result.metadata["score"] = -score
                 
                 # Count only successful crawls toward max_pages limit
                 if result.success:
                     self._pages_crawled += 1
+                    # Check if we've reached the limit during batch processing
+                    if self._pages_crawled >= self.max_pages:
+                        self.logger.info(f"Max pages limit ({self.max_pages}) reached during batch, stopping crawl")
+                        break  # Exit the generator
                 
                 yield result
                 
@@ -196,7 +204,7 @@ class BestFirstCrawlingStrategy(DeepCrawlStrategy):
                     for new_url, new_parent in new_links:
                         new_depth = depths.get(new_url, depth + 1)
                         new_score = self.url_scorer.score(new_url) if self.url_scorer else 0
-                        await queue.put((new_score, new_depth, new_url, new_parent))
+                        await queue.put((-new_score, new_depth, new_url, new_parent))
 
         # End of crawl.
 

crawl4ai/deep_crawling/bfs_strategy.py

@@ -157,6 +157,11 @@ class BFSDeepCrawlStrategy(DeepCrawlStrategy):
         results: List[CrawlResult] = []
 
         while current_level and not self._cancel_event.is_set():
+            # Check if we've already reached max_pages before starting a new level
+            if self._pages_crawled >= self.max_pages:
+                self.logger.info(f"Max pages limit ({self.max_pages}) reached, stopping crawl")
+                break
+            
             next_level: List[Tuple[str, Optional[str]]] = []
             urls = [url for url, _ in current_level]
 
@@ -221,6 +226,10 @@ class BFSDeepCrawlStrategy(DeepCrawlStrategy):
                 # Count only successful crawls
                 if result.success:
                     self._pages_crawled += 1
+                    # Check if we've reached the limit during batch processing
+                    if self._pages_crawled >= self.max_pages:
+                        self.logger.info(f"Max pages limit ({self.max_pages}) reached during batch, stopping crawl")
+                        break  # Exit the generator
                 
                 results_count += 1
                 yield result

crawl4ai/deep_crawling/dfs_strategy.py

@@ -49,6 +49,10 @@ class DFSDeepCrawlStrategy(BFSDeepCrawlStrategy):
                 # Count only successful crawls toward max_pages limit
                 if result.success:
                     self._pages_crawled += 1
+                    # Check if we've reached the limit during batch processing
+                    if self._pages_crawled >= self.max_pages:
+                        self.logger.info(f"Max pages limit ({self.max_pages}) reached during batch, stopping crawl")
+                        break  # Exit the generator
                     
                     # Only discover links from successful crawls
                     new_links: List[Tuple[str, Optional[str]]] = []
@@ -94,6 +98,10 @@ class DFSDeepCrawlStrategy(BFSDeepCrawlStrategy):
                 # and only discover links from successful crawls
                 if result.success:
                     self._pages_crawled += 1
+                    # Check if we've reached the limit during batch processing
+                    if self._pages_crawled >= self.max_pages:
+                        self.logger.info(f"Max pages limit ({self.max_pages}) reached during batch, stopping crawl")
+                        break  # Exit the generator
                     
                     new_links: List[Tuple[str, Optional[str]]] = []
                     await self.link_discovery(result, url, depth, visited, new_links, depths)

crawl4ai/deep_crawling/filters.py

@@ -227,10 +227,21 @@ class URLPatternFilter(URLFilter):
         # Prefix check (/foo/*)
         if self._simple_prefixes:
             path = url.split("?")[0]
-            if any(path.startswith(p) for p in self._simple_prefixes):
-                result = True
-                self._update_stats(result)
-                return not result if self._reverse else result
+            # if any(path.startswith(p) for p in self._simple_prefixes):
+            #     result = True
+            #     self._update_stats(result)
+            #     return not result if self._reverse else result
+            ####
+            # Modified the prefix matching logic to ensure path boundary checking:
+            # - Check if the matched prefix is followed by a path separator (`/`), query parameter (`?`), fragment (`#`), or is at the end of the path
+            # - This ensures `/api/` only matches complete path segments, not substrings like `/apiv2/`
+            ####
+            for prefix in self._simple_prefixes:
+                if path.startswith(prefix):
+                    if len(path) == len(prefix) or path[len(prefix)] in ['/', '?', '#']:
+                        result = True
+                        self._update_stats(result)
+                        return not result if self._reverse else result
 
         # Complex patterns
         if self._path_patterns:
@@ -337,6 +348,15 @@ class ContentTypeFilter(URLFilter):
         "sqlite": "application/vnd.sqlite3",
         # Placeholder
         "unknown": "application/octet-stream",  # Fallback for unknown file types
+        # php
+        "php": "application/x-httpd-php",
+        "php3": "application/x-httpd-php",
+        "php4": "application/x-httpd-php",
+        "php5": "application/x-httpd-php",
+        "php7": "application/x-httpd-php",
+        "phtml": "application/x-httpd-php",
+        "phps": "application/x-httpd-php-source",
+
     }
 
     @staticmethod

crawl4ai/docker_client.py

@@ -73,6 +73,8 @@ class Crawl4aiDockerClient:
     def _prepare_request(self, urls: List[str], browser_config: Optional[BrowserConfig] = None, 
                        crawler_config: Optional[CrawlerRunConfig] = None) -> Dict[str, Any]:
         """Prepare request data from configs."""
+        if self._token:
+            self._http_client.headers["Authorization"] = f"Bearer {self._token}"
         return {
             "urls": urls,
             "browser_config": browser_config.dump() if browser_config else {},
@@ -103,8 +105,6 @@ class Crawl4aiDockerClient:
         crawler_config: Optional[CrawlerRunConfig] = None
     ) -> Union[CrawlResult, List[CrawlResult], AsyncGenerator[CrawlResult, None]]:
         """Execute a crawl operation."""
-        if not self._token:
-            raise Crawl4aiClientError("Authentication required. Call authenticate() first.")
         await self._check_server()
         
         data = self._prepare_request(urls, browser_config, crawler_config)
@@ -140,8 +140,6 @@ class Crawl4aiDockerClient:
 
     async def get_schema(self) -> Dict[str, Any]:
         """Retrieve configuration schemas."""
-        if not self._token:
-            raise Crawl4aiClientError("Authentication required. Call authenticate() first.")
         response = await self._request("GET", "/schema")
         return response.json()
 
@@ -167,4 +165,4 @@ async def main():
         print(schema)
 
 if __name__ == "__main__":
-    asyncio.run(main())
+    asyncio.run(main())

crawl4ai/extraction_strategy.py

@@ -656,11 +656,11 @@ class LLMExtractionStrategy(ExtractionStrategy):
             self.total_usage.total_tokens += usage.total_tokens
 
             try:
-                response = response.choices[0].message.content
+                content = response.choices[0].message.content
                 blocks = None
 
                 if self.force_json_response:
-                    blocks = json.loads(response)
+                    blocks = json.loads(content)
                     if isinstance(blocks, dict):
                         # If it has only one key which calue is list then assign that to blocks, exampled: {"news": [..]}
                         if len(blocks) == 1 and isinstance(list(blocks.values())[0], list):
@@ -673,7 +673,7 @@ class LLMExtractionStrategy(ExtractionStrategy):
                         blocks = blocks
                 else: 
                     # blocks = extract_xml_data(["blocks"], response.choices[0].message.content)["blocks"]
-                    blocks = extract_xml_data(["blocks"], response)["blocks"]
+                    blocks = extract_xml_data(["blocks"], content)["blocks"]
                     blocks = json.loads(blocks)
 
                 for block in blocks:

crawl4ai/install.py

@@ -119,6 +119,32 @@ def install_playwright():
         logger.warning(
             f"Please run '{sys.executable} -m playwright install --with-deps' manually after the installation."
         )
+    
+    # Install Patchright browsers for undetected browser support
+    logger.info("Installing Patchright browsers for undetected mode...", tag="INIT")
+    try:
+        subprocess.check_call(
+            [
+                sys.executable,
+                "-m",
+                "patchright",
+                "install",
+                "--with-deps",
+                "--force",
+                "chromium",
+            ]
+        )
+        logger.success(
+            "Patchright installation completed successfully.", tag="COMPLETE"
+        )
+    except subprocess.CalledProcessError:
+        logger.warning(
+            f"Please run '{sys.executable} -m patchright install --with-deps' manually after the installation."
+        )
+    except Exception:
+        logger.warning(
+            f"Please run '{sys.executable} -m patchright install --with-deps' manually after the installation."
+        )
 
 
 def run_migration():

crawl4ai/legacy/web_crawler.py

@@ -11,7 +11,7 @@ from .extraction_strategy import *
 from .crawler_strategy import *
 from typing import List
 from concurrent.futures import ThreadPoolExecutor
-from .content_scraping_strategy import WebScrapingStrategy
+from ..content_scraping_strategy import LXMLWebScrapingStrategy as WebScrapingStrategy
 from .config import *
 import warnings
 import json

crawl4ai/memory_utils.py

@@ -0,0 +1,79 @@
+import psutil
+import platform
+import subprocess
+from typing import Tuple
+
+
+def get_true_available_memory_gb() -> float:
+    """Get truly available memory including inactive pages (cross-platform)"""
+    vm = psutil.virtual_memory()
+
+    if platform.system() == 'Darwin':  # macOS
+        # On macOS, we need to include inactive memory too
+        try:
+            # Use vm_stat to get accurate values
+            result = subprocess.run(['vm_stat'], capture_output=True, text=True)
+            lines = result.stdout.split('\n')
+
+            page_size = 16384  # macOS page size
+            pages = {}
+
+            for line in lines:
+                if 'Pages free:' in line:
+                    pages['free'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages inactive:' in line:
+                    pages['inactive'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages speculative:' in line:
+                    pages['speculative'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages purgeable:' in line:
+                    pages['purgeable'] = int(line.split()[-1].rstrip('.'))
+
+            # Calculate total available (free + inactive + speculative + purgeable)
+            total_available_pages = (
+                pages.get('free', 0) + 
+                pages.get('inactive', 0) + 
+                pages.get('speculative', 0) + 
+                pages.get('purgeable', 0)
+            )
+            available_gb = (total_available_pages * page_size) / (1024**3)
+
+            return available_gb
+        except:
+            # Fallback to psutil
+            return vm.available / (1024**3)
+    else:
+        # For Windows and Linux, psutil.available is accurate
+        return vm.available / (1024**3)
+
+
+def get_true_memory_usage_percent() -> float:
+    """
+    Get memory usage percentage that accounts for platform differences.
+    
+    Returns:
+        float: Memory usage percentage (0-100)
+    """
+    vm = psutil.virtual_memory()
+    total_gb = vm.total / (1024**3)
+    available_gb = get_true_available_memory_gb()
+    
+    # Calculate used percentage based on truly available memory
+    used_percent = 100.0 * (total_gb - available_gb) / total_gb
+    
+    # Ensure it's within valid range
+    return max(0.0, min(100.0, used_percent))
+
+
+def get_memory_stats() -> Tuple[float, float, float]:
+    """
+    Get comprehensive memory statistics.
+    
+    Returns:
+        Tuple[float, float, float]: (used_percent, available_gb, total_gb)
+    """
+    vm = psutil.virtual_memory()
+    total_gb = vm.total / (1024**3)
+    available_gb = get_true_available_memory_gb()
+    used_percent = get_true_memory_usage_percent()
+    
+    return used_percent, available_gb, total_gb

crawl4ai/prompts.py

@@ -1056,7 +1056,7 @@ Your output must:
 </output_requirements>
 """
 
-GENERATE_SCRIPT_PROMPT = """You are a world-class browser automation specialist. Your sole purpose is to convert a natural language objective and a snippet of HTML into the most **efficient, robust, and simple** script possible to prepare a web page for data extraction.
+GENERATE_SCRIPT_PROMPT = r"""You are a world-class browser automation specialist. Your sole purpose is to convert a natural language objective and a snippet of HTML into the most **efficient, robust, and simple** script possible to prepare a web page for data extraction.
 
 Your scripts run **before the crawl** to handle dynamic content, user interactions, and other obstacles. You are a master of two tools: raw **JavaScript** and the high-level **Crawl4ai Script (c4a)**.
 

crawl4ai/types.py

@@ -23,8 +23,9 @@ SeedingConfig = Union['SeedingConfigType']
 
 # Content scraping types
 ContentScrapingStrategy = Union['ContentScrapingStrategyType']
-WebScrapingStrategy = Union['WebScrapingStrategyType']
 LXMLWebScrapingStrategy = Union['LXMLWebScrapingStrategyType']
+# Backward compatibility alias
+WebScrapingStrategy = Union['LXMLWebScrapingStrategyType']
 
 # Proxy types
 ProxyRotationStrategy = Union['ProxyRotationStrategyType']
@@ -114,7 +115,6 @@ if TYPE_CHECKING:
     # Content scraping imports
     from .content_scraping_strategy import (
         ContentScrapingStrategy as ContentScrapingStrategyType,
-        WebScrapingStrategy as WebScrapingStrategyType,
         LXMLWebScrapingStrategy as LXMLWebScrapingStrategyType,
     )
     

crawl4ai/utils.py

@@ -32,7 +32,6 @@ import hashlib
 
 from urllib.robotparser import RobotFileParser
 import aiohttp
-from urllib.parse import urlparse, urlunparse
 from functools import lru_cache
 
 from packaging import version
@@ -43,6 +42,37 @@ from itertools import chain
 from collections import deque
 from typing import  Generator, Iterable
 
+import numpy as np
+
+from urllib.parse import (
+    urljoin, urlparse, urlunparse,
+    parse_qsl, urlencode, quote, unquote
+)
+
+
+# Monkey patch to fix wildcard handling in urllib.robotparser
+from urllib.robotparser import RuleLine
+import re
+
+original_applies_to = RuleLine.applies_to
+
+def patched_applies_to(self, filename):
+   # Handle wildcards in paths
+   if '*' in self.path or '%2A' in self.path or self.path in ("*", "%2A"):
+       pattern = self.path.replace('%2A', '*')
+       pattern = re.escape(pattern).replace('\\*', '.*')
+       pattern = '^' + pattern
+       if pattern.endswith('\\$'):
+           pattern = pattern[:-2] + '$'
+       try:
+           return bool(re.match(pattern, filename))
+       except re.error:
+           return original_applies_to(self, filename)
+   return original_applies_to(self, filename)
+
+RuleLine.applies_to = patched_applies_to
+# Monkey patch ends
+
 def chunk_documents(
     documents: Iterable[str],
     chunk_token_threshold: int,
@@ -311,7 +341,7 @@ class RobotsParser:
                 robots_url = f"{scheme}://{domain}/robots.txt"
                 
                 async with aiohttp.ClientSession() as session:
-                    async with session.get(robots_url, timeout=2) as response:
+                    async with session.get(robots_url, timeout=2, ssl=False) as response:
                         if response.status == 200:
                             rules = await response.text()
                             self._cache_rules(domain, rules)
@@ -1487,8 +1517,29 @@ def extract_metadata_using_lxml(html, doc=None):
     head = head[0]
 
     # Title - using XPath
+    # title = head.xpath(".//title/text()")
+    # metadata["title"] = title[0].strip() if title else None
+
+    # === Title Extraction - New Approach ===
+    # Attempt to extract <title> using XPath
     title = head.xpath(".//title/text()")
-    metadata["title"] = title[0].strip() if title else None
+    title = title[0] if title else None
+
+    # Fallback: Use .find() in case XPath fails due to malformed HTML
+    if not title:
+        title_el = doc.find(".//title")
+        title = title_el.text if title_el is not None else None
+
+    # Final fallback: Use OpenGraph or Twitter title if <title> is missing or empty
+    if not title:
+        title_candidates = (
+            doc.xpath("//meta[@property='og:title']/@content") or
+            doc.xpath("//meta[@name='twitter:title']/@content")
+        )
+        title = title_candidates[0] if title_candidates else None
+
+    # Strip and assign title
+    metadata["title"] = title.strip() if title else None
 
     # Meta description - using XPath with multiple attribute conditions
     description = head.xpath('.//meta[@name="description"]/@content')
@@ -1517,6 +1568,14 @@ def extract_metadata_using_lxml(html, doc=None):
         content = tag.get("content", "").strip()
         if property_name and content:
             metadata[property_name] = content
+   
+   # Article metadata
+    article_tags = head.xpath('.//meta[starts-with(@property, "article:")]')
+    for tag in article_tags:
+        property_name = tag.get("property", "").strip()
+        content = tag.get("content", "").strip()
+        if property_name and content:
+            metadata[property_name] = content
 
     return metadata
 
@@ -1592,7 +1651,15 @@ def extract_metadata(html, soup=None):
         content = tag.get("content", "").strip()
         if property_name and content:
             metadata[property_name] = content
-
+    
+    # Article metadata
+    article_tags = head.find_all("meta", attrs={"property": re.compile(r"^article:")})
+    for tag in article_tags:
+        property_name = tag.get("property", "").strip()
+        content = tag.get("content", "").strip()
+        if property_name and content:
+            metadata[property_name] = content
+    
     return metadata
 
 
@@ -2061,13 +2128,101 @@ def normalize_url(href, base_url):
     parsed_base = urlparse(base_url)
     if not parsed_base.scheme or not parsed_base.netloc:
         raise ValueError(f"Invalid base URL format: {base_url}")
-
-    # Ensure base_url ends with a trailing slash if it's a directory path
-    if not base_url.endswith('/'):
-        base_url = base_url + '/'
+    
+    if  parsed_base.scheme.lower() not in ["http", "https"]:
+        # Handle special protocols
+        raise ValueError(f"Invalid base URL format: {base_url}")
+    cleaned_href = href.strip()
 
     # Use urljoin to handle all cases
-    normalized = urljoin(base_url, href.strip())
+    return urljoin(base_url, cleaned_href)
+
+
+
+
+def normalize_url(
+    href: str,
+    base_url: str,
+    *,
+    drop_query_tracking=True,
+    sort_query=True,
+    keep_fragment=False,
+    extra_drop_params=None
+):
+    """
+    Extended URL normalizer
+
+    Parameters
+    ----------
+    href : str
+        The raw link extracted from a page.
+    base_url : str
+        The page’s canonical URL (used to resolve relative links).
+    drop_query_tracking : bool (default True)
+        Remove common tracking query parameters.
+    sort_query : bool (default True)
+        Alphabetically sort query keys for deterministic output.
+    keep_fragment : bool (default False)
+        Preserve the hash fragment (#section) if you need in-page links.
+    extra_drop_params : Iterable[str] | None
+        Additional query keys to strip (case-insensitive).
+
+    Returns
+    -------
+    str | None
+        A clean, canonical URL or None if href is empty/None.
+    """
+    if not href:
+        return None
+
+    # Resolve relative paths first
+    full_url = urljoin(base_url, href.strip())
+
+    # Parse once, edit parts, then rebuild
+    parsed = urlparse(full_url)
+
+    # ── netloc ──
+    netloc = parsed.netloc.lower()
+
+    # ── path ──
+    # Strip duplicate slashes and trailing “/” (except root)
+    path = quote(unquote(parsed.path))
+    if path.endswith('/') and path != '/':
+        path = path.rstrip('/')
+
+    # ── query ──
+    query = parsed.query
+    if query:
+        # explode, mutate, then rebuild
+        params = [(k.lower(), v) for k, v in parse_qsl(query, keep_blank_values=True)]
+
+        if drop_query_tracking:
+            default_tracking = {
+                'utm_source', 'utm_medium', 'utm_campaign', 'utm_term',
+                'utm_content', 'gclid', 'fbclid', 'ref', 'ref_src'
+            }
+            if extra_drop_params:
+                default_tracking |= {p.lower() for p in extra_drop_params}
+            params = [(k, v) for k, v in params if k not in default_tracking]
+
+        if sort_query:
+            params.sort(key=lambda kv: kv[0])
+
+        query = urlencode(params, doseq=True) if params else ''
+
+    # ── fragment ──
+    fragment = parsed.fragment if keep_fragment else ''
+
+    # Re-assemble
+    normalized = urlunparse((
+        parsed.scheme,
+        netloc,
+        path,
+        parsed.params,
+        query,
+        fragment
+    ))
+
     return normalized
 
 
@@ -3148,3 +3303,114 @@ def calculate_total_score(
     
     return max(0.0, min(total, 10.0))
 
+
+# Embedding utilities
+async def get_text_embeddings(
+    texts: List[str], 
+    llm_config: Optional[Dict] = None,
+    model_name: str = "sentence-transformers/all-MiniLM-L6-v2",
+    batch_size: int = 32
+) -> np.ndarray:
+    """
+    Compute embeddings for a list of texts using specified model.
+    
+    Args:
+        texts: List of texts to embed
+        llm_config: Optional LLM configuration for API-based embeddings
+        model_name: Model name (used when llm_config is None)
+        batch_size: Batch size for processing
+        
+    Returns:
+        numpy array of embeddings
+    """
+    import numpy as np
+    
+    if not texts:
+        return np.array([])
+    
+    # If LLMConfig provided, use litellm for embeddings
+    if llm_config is not None:
+        from litellm import aembedding
+        
+        # Get embedding model from config or use default
+        embedding_model = llm_config.get('provider', 'text-embedding-3-small')
+        api_base = llm_config.get('base_url', llm_config.get('api_base'))
+        
+        # Prepare kwargs
+        kwargs = {
+            'model': embedding_model,
+            'input': texts,
+            'api_key': llm_config.get('api_token', llm_config.get('api_key'))
+        }
+        
+        if api_base:
+            kwargs['api_base'] = api_base
+            
+        # Handle OpenAI-compatible endpoints
+        if api_base and 'openai/' not in embedding_model:
+            kwargs['model'] = f"openai/{embedding_model}"
+        
+        # Get embeddings
+        response = await aembedding(**kwargs)
+        
+        # Extract embeddings from response
+        embeddings = []
+        for item in response.data:
+            embeddings.append(item['embedding'])
+            
+        return np.array(embeddings)
+    
+    # Default: use sentence-transformers
+    else:
+        # Lazy load to avoid importing heavy libraries unless needed
+        try:
+            from sentence_transformers import SentenceTransformer
+        except ImportError:
+            raise ImportError(
+                "sentence-transformers is required for local embeddings. "
+                "Install it with: pip install 'crawl4ai[transformer]' or pip install sentence-transformers"
+            )
+        
+        # Cache the model in function attribute to avoid reloading
+        if not hasattr(get_text_embeddings, '_models'):
+            get_text_embeddings._models = {}
+        
+        if model_name not in get_text_embeddings._models:
+            get_text_embeddings._models[model_name] = SentenceTransformer(model_name)
+        
+        encoder = get_text_embeddings._models[model_name]
+        
+        # Batch encode for efficiency
+        embeddings = encoder.encode(
+            texts,
+            batch_size=batch_size,
+            show_progress_bar=False,
+            convert_to_numpy=True
+        )
+        
+        return embeddings
+
+
+def get_text_embeddings_sync(
+    texts: List[str],
+    llm_config: Optional[Dict] = None,
+    model_name: str = "sentence-transformers/all-MiniLM-L6-v2",
+    batch_size: int = 32
+) -> np.ndarray:
+    """Synchronous wrapper for get_text_embeddings"""
+    import numpy as np
+    return asyncio.run(get_text_embeddings(texts, llm_config, model_name, batch_size))
+
+
+def cosine_similarity(vec1: np.ndarray, vec2: np.ndarray) -> float:
+    """Calculate cosine similarity between two vectors"""
+    import numpy as np
+    dot_product = np.dot(vec1, vec2)
+    norm_product = np.linalg.norm(vec1) * np.linalg.norm(vec2)
+    return float(dot_product / norm_product) if norm_product != 0 else 0.0
+
+
+def cosine_distance(vec1: np.ndarray, vec2: np.ndarray) -> float:
+    """Calculate cosine distance (1 - similarity) between two vectors"""
+    return 1 - cosine_similarity(vec1, vec2)
+

deploy/docker/.llm.env.example

@@ -5,4 +5,9 @@ ANTHROPIC_API_KEY=your_anthropic_key_here
 GROQ_API_KEY=your_groq_key_here
 TOGETHER_API_KEY=your_together_key_here
 MISTRAL_API_KEY=your_mistral_key_here
-GEMINI_API_TOKEN=your_gemini_key_here
+GEMINI_API_TOKEN=your_gemini_key_here
+
+# Optional: Override the default LLM provider
+# Examples: "openai/gpt-4", "anthropic/claude-3-opus", "deepseek/chat", etc.
+# If not set, uses the provider specified in config.yml (default: openai/gpt-4o-mini)
+# LLM_PROVIDER=anthropic/claude-3-opus

deploy/docker/README.md

@@ -58,13 +58,15 @@ Pull and run images directly from Docker Hub without building locally.
 
 #### 1. Pull the Image
 
-Our latest release candidate is `0.6.0-r1`. Images are built with multi-arch manifests, so Docker automatically pulls the correct version for your system.
+Our latest release candidate is `0.7.0-r1`. Images are built with multi-arch manifests, so Docker automatically pulls the correct version for your system.
+
+> ⚠️ **Important Note**: The `latest` tag currently points to the stable `0.6.0` version. After testing and validation, `0.7.0` (without -r1) will be released and `latest` will be updated. For now, please use `0.7.0-r1` to test the new features.
 
 ```bash
-# Pull the release candidate (recommended for latest features)
-docker pull unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number
+# Pull the release candidate (for testing new features)
+docker pull unclecode/crawl4ai:0.7.0-r1
 
-# Or pull the latest stable version
+# Or pull the current stable version (0.6.0)
 docker pull unclecode/crawl4ai:latest
 ```
 
@@ -99,7 +101,7 @@ EOL
       -p 11235:11235 \
       --name crawl4ai \
       --shm-size=1g \
-      unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number
+      unclecode/crawl4ai:0.7.0-r1
     ```
 
 *   **With LLM support:**
@@ -110,7 +112,7 @@ EOL
       --name crawl4ai \
       --env-file .llm.env \
       --shm-size=1g \
-      unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number
+      unclecode/crawl4ai:0.7.0-r1
     ```
 
 > The server will be available at `http://localhost:11235`. Visit `/playground` to access the interactive testing interface.
@@ -124,7 +126,7 @@ docker stop crawl4ai && docker rm crawl4ai
 #### Docker Hub Versioning Explained
 
 *   **Image Name:** `unclecode/crawl4ai`
-*   **Tag Format:** `LIBRARY_VERSION[-SUFFIX]` (e.g., `0.6.0-r1`)
+*   **Tag Format:** `LIBRARY_VERSION[-SUFFIX]` (e.g., `0.7.0-r1`)
     *   `LIBRARY_VERSION`: The semantic version of the core `crawl4ai` Python library
     *   `SUFFIX`: Optional tag for release candidates (``) and revisions (`r1`)
 *   **`latest` Tag:** Points to the most recent stable version
@@ -152,6 +154,29 @@ cp deploy/docker/.llm.env.example .llm.env
 # Now edit .llm.env and add your API keys
 ```
 
+**Flexible LLM Provider Configuration:**
+
+The Docker setup now supports flexible LLM provider configuration through three methods:
+
+1. **Environment Variable** (Highest Priority): Set `LLM_PROVIDER` to override the default
+   ```bash
+   export LLM_PROVIDER="anthropic/claude-3-opus"
+   # Or in your .llm.env file:
+   # LLM_PROVIDER=anthropic/claude-3-opus
+   ```
+
+2. **API Request Parameter**: Specify provider per request
+   ```json
+   {
+     "url": "https://example.com",
+     "provider": "groq/mixtral-8x7b"
+   }
+   ```
+
+3. **Config File Default**: Falls back to `config.yml` (default: `openai/gpt-4o-mini`)
+
+The system automatically selects the appropriate API key based on the provider.
+
 #### 3. Build and Run with Compose
 
 The `docker-compose.yml` file in the project root provides a simplified approach that automatically handles architecture detection using buildx.
@@ -160,7 +185,7 @@ The `docker-compose.yml` file in the project root provides a simplified approach
     ```bash
     # Pulls and runs the release candidate from Docker Hub
     # Automatically selects the correct architecture
-    IMAGE=unclecode/crawl4ai:0.6.0-rN # Use your favorite revision number docker compose up -d
+    IMAGE=unclecode/crawl4ai:0.7.0-r1 docker compose up -d
     ```
 
 *   **Build and Run Locally:**
@@ -666,7 +691,7 @@ app:
 
 # Default LLM Configuration
 llm:
-  provider: "openai/gpt-4o-mini"
+  provider: "openai/gpt-4o-mini"  # Can be overridden by LLM_PROVIDER env var
   api_key_env: "OPENAI_API_KEY"
   # api_key: sk-...  # If you pass the API key directly then api_key_env will be ignored
 

deploy/docker/api.py

@@ -5,6 +5,7 @@ from typing import List, Tuple, Dict
 from functools import partial
 from uuid import uuid4
 from datetime import datetime
+from base64 import b64encode
 
 import logging
 from typing import Optional, AsyncGenerator
@@ -39,7 +40,9 @@ from utils import (
     get_base_url,
     is_task_id,
     should_cleanup_task,
-    decode_redis_hash
+    decode_redis_hash,
+    get_llm_api_key,
+    validate_llm_provider
 )
 
 import psutil, time
@@ -88,10 +91,12 @@ async def handle_llm_qa(
 
     Answer:"""
 
+        # api_token=os.environ.get(config["llm"].get("api_key_env", ""))
+
         response = perform_completion_with_backoff(
             provider=config["llm"]["provider"],
             prompt_with_variables=prompt,
-            api_token=os.environ.get(config["llm"].get("api_key_env", ""))
+            api_token=get_llm_api_key(config)
         )
 
         return response.choices[0].message.content
@@ -109,19 +114,23 @@ async def process_llm_extraction(
     url: str,
     instruction: str,
     schema: Optional[str] = None,
-    cache: str = "0"
+    cache: str = "0",
+    provider: Optional[str] = None
 ) -> None:
     """Process LLM extraction in background."""
     try:
-        # If config['llm'] has api_key then ignore the api_key_env
-        api_key = ""
-        if "api_key" in config["llm"]:
-            api_key = config["llm"]["api_key"]
-        else:
-            api_key = os.environ.get(config["llm"].get("api_key_env", None), "")
+        # Validate provider
+        is_valid, error_msg = validate_llm_provider(config, provider)
+        if not is_valid:
+            await redis.hset(f"task:{task_id}", mapping={
+                "status": TaskStatus.FAILED,
+                "error": error_msg
+            })
+            return
+        api_key = get_llm_api_key(config, provider)
         llm_strategy = LLMExtractionStrategy(
             llm_config=LLMConfig(
-                provider=config["llm"]["provider"],
+                provider=provider or config["llm"]["provider"],
                 api_token=api_key
             ),
             instruction=instruction,
@@ -168,10 +177,19 @@ async def handle_markdown_request(
     filter_type: FilterType,
     query: Optional[str] = None,
     cache: str = "0",
-    config: Optional[dict] = None
+    config: Optional[dict] = None,
+    provider: Optional[str] = None
 ) -> str:
     """Handle markdown generation requests."""
     try:
+        # Validate provider if using LLM filter
+        if filter_type == FilterType.LLM:
+            is_valid, error_msg = validate_llm_provider(config, provider)
+            if not is_valid:
+                raise HTTPException(
+                    status_code=status.HTTP_400_BAD_REQUEST,
+                    detail=error_msg
+                )
         decoded_url = unquote(url)
         if not decoded_url.startswith(('http://', 'https://')):
             decoded_url = 'https://' + decoded_url
@@ -184,8 +202,8 @@ async def handle_markdown_request(
                 FilterType.BM25: BM25ContentFilter(user_query=query or ""),
                 FilterType.LLM: LLMContentFilter(
                     llm_config=LLMConfig(
-                        provider=config["llm"]["provider"],
-                        api_token=os.environ.get(config["llm"].get("api_key_env", None), ""),
+                        provider=provider or config["llm"]["provider"],
+                        api_token=get_llm_api_key(config, provider),
                     ),
                     instruction=query or "Extract main content"
                 )
@@ -229,7 +247,8 @@ async def handle_llm_request(
     query: Optional[str] = None,
     schema: Optional[str] = None,
     cache: str = "0",
-    config: Optional[dict] = None
+    config: Optional[dict] = None,
+    provider: Optional[str] = None
 ) -> JSONResponse:
     """Handle LLM extraction requests."""
     base_url = get_base_url(request)
@@ -259,7 +278,8 @@ async def handle_llm_request(
             schema,
             cache,
             base_url,
-            config
+            config,
+            provider
         )
 
     except Exception as e:
@@ -303,7 +323,8 @@ async def create_new_task(
     schema: Optional[str],
     cache: str,
     base_url: str,
-    config: dict
+    config: dict,
+    provider: Optional[str] = None
 ) -> JSONResponse:
     """Create and initialize a new task."""
     decoded_url = unquote(input_path)
@@ -327,7 +348,8 @@ async def create_new_task(
         decoded_url,
         query,
         schema,
-        cache
+        cache,
+        provider
     )
 
     return JSONResponse({
@@ -371,6 +393,9 @@ async def stream_results(crawler: AsyncWebCrawler, results_gen: AsyncGenerator)
                 server_memory_mb = _get_memory_mb()
                 result_dict = result.model_dump()
                 result_dict['server_memory_mb'] = server_memory_mb
+                # If PDF exists, encode it to base64
+                if result_dict.get('pdf') is not None:
+                    result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
                 logger.info(f"Streaming result for {result_dict.get('url', 'unknown')}")
                 data = json.dumps(result_dict, default=datetime_handler) + "\n"
                 yield data.encode('utf-8')
@@ -443,10 +468,19 @@ async def handle_crawl_request(
             mem_delta_mb = end_mem_mb - start_mem_mb # <--- Calculate delta
             peak_mem_mb = max(peak_mem_mb if peak_mem_mb else 0, end_mem_mb) # <--- Get peak memory
         logger.info(f"Memory usage: Start: {start_mem_mb} MB, End: {end_mem_mb} MB, Delta: {mem_delta_mb} MB, Peak: {peak_mem_mb} MB")
-                              
+
+        # Process results to handle PDF bytes
+        processed_results = []
+        for result in results:
+            result_dict = result.model_dump()
+            # If PDF exists, encode it to base64
+            if result_dict.get('pdf') is not None:
+                result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
+            processed_results.append(result_dict)
+            
         return {
             "success": True,
-            "results": [result.model_dump() for result in results],
+            "results": processed_results,
             "server_processing_time_s": end_time - start_time,
             "server_memory_delta_mb": mem_delta_mb,
             "server_peak_memory_mb": peak_mem_mb
@@ -459,7 +493,7 @@ async def handle_crawl_request(
             #      await crawler.close()
             #  except Exception as close_e:
             #       logger.error(f"Error closing crawler during exception handling: {close_e}")
-            logger.error(f"Error closing crawler during exception handling: {close_e}")
+            logger.error(f"Error closing crawler during exception handling: {str(e)}")
 
         # Measure memory even on error if possible
         end_mem_mb_error = _get_memory_mb()
@@ -518,7 +552,7 @@ async def handle_stream_crawl_request(
             #       await crawler.close()
             #  except Exception as close_e:
             #       logger.error(f"Error closing crawler during stream setup exception: {close_e}")
-            logger.error(f"Error closing crawler during stream setup exception: {close_e}")
+            logger.error(f"Error closing crawler during stream setup exception: {str(e)}")
         logger.error(f"Stream crawl error: {str(e)}", exc_info=True)
         # Raising HTTPException here will prevent streaming response
         raise HTTPException(

deploy/docker/c4ai-doc-context.md

@@ -332,7 +332,7 @@ The `clone()` method:
 ### Key fields to note
 
 1. **`provider`**:  
-- Which LLM provoder to use. 
+- Which LLM provider to use. 
 - Possible values are `"ollama/llama3","groq/llama3-70b-8192","groq/llama3-8b-8192", "openai/gpt-4o-mini" ,"openai/gpt-4o","openai/o1-mini","openai/o1-preview","openai/o3-mini","openai/o3-mini-high","anthropic/claude-3-haiku-20240307","anthropic/claude-3-opus-20240229","anthropic/claude-3-sonnet-20240229","anthropic/claude-3-5-sonnet-20240620","gemini/gemini-pro","gemini/gemini-1.5-pro","gemini/gemini-2.0-flash","gemini/gemini-2.0-flash-exp","gemini/gemini-2.0-flash-lite-preview-02-05","deepseek/deepseek-chat"`<br/>*(default: `"openai/gpt-4o-mini"`)*
 
 2. **`api_token`**:  
@@ -403,7 +403,7 @@ async def main():
 
     md_generator = DefaultMarkdownGenerator(
     content_filter=filter,
-    options={"ignore_links": True}
+    options={"ignore_links": True})
 
     # 4) Crawler run config: skip cache, use extraction
     run_conf = CrawlerRunConfig(
@@ -3760,11 +3760,11 @@ To crawl a live web page, provide the URL starting with `http://` or `https://`,
 
 ```python
 import asyncio
-from crawl4ai import AsyncWebCrawler
+from crawl4ai import AsyncWebCrawler, CacheMode
 from crawl4ai.async_configs import CrawlerRunConfig
 
 async def crawl_web():
-    config = CrawlerRunConfig(bypass_cache=True)
+    config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(
             url="https://en.wikipedia.org/wiki/apple", 
@@ -3785,13 +3785,13 @@ To crawl a local HTML file, prefix the file path with `file://`.
 
 ```python
 import asyncio
-from crawl4ai import AsyncWebCrawler
+from crawl4ai import AsyncWebCrawler, CacheMode
 from crawl4ai.async_configs import CrawlerRunConfig
 
 async def crawl_local_file():
     local_file_path = "/path/to/apple.html"  # Replace with your file path
     file_url = f"file://{local_file_path}"
-    config = CrawlerRunConfig(bypass_cache=True)
+    config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
     
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(url=file_url, config=config)
@@ -3810,13 +3810,13 @@ To crawl raw HTML content, prefix the HTML string with `raw:`.
 
 ```python
 import asyncio
-from crawl4ai import AsyncWebCrawler
+from crawl4ai import AsyncWebCrawler, CacheMode
 from crawl4ai.async_configs import CrawlerRunConfig
 
 async def crawl_raw_html():
     raw_html = "<html><body><h1>Hello, World!</h1></body></html>"
     raw_html_url = f"raw:{raw_html}"
-    config = CrawlerRunConfig(bypass_cache=True)
+    config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
     
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(url=raw_html_url, config=config)
@@ -3845,7 +3845,7 @@ import os
 import sys
 import asyncio
 from pathlib import Path
-from crawl4ai import AsyncWebCrawler
+from crawl4ai import AsyncWebCrawler, CacheMode
 from crawl4ai.async_configs import CrawlerRunConfig
 
 async def main():
@@ -3856,7 +3856,7 @@ async def main():
     async with AsyncWebCrawler() as crawler:
         # Step 1: Crawl the Web URL
         print("\n=== Step 1: Crawling the Wikipedia URL ===")
-        web_config = CrawlerRunConfig(bypass_cache=True)
+        web_config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
         result = await crawler.arun(url=wikipedia_url, config=web_config)
 
         if not result.success:
@@ -3871,7 +3871,7 @@ async def main():
         # Step 2: Crawl from the Local HTML File
         print("=== Step 2: Crawling from the Local HTML File ===")
         file_url = f"file://{html_file_path.resolve()}"
-        file_config = CrawlerRunConfig(bypass_cache=True)
+        file_config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
         local_result = await crawler.arun(url=file_url, config=file_config)
 
         if not local_result.success:
@@ -3887,7 +3887,7 @@ async def main():
         with open(html_file_path, 'r', encoding='utf-8') as f:
             raw_html_content = f.read()
         raw_html_url = f"raw:{raw_html_content}"
-        raw_config = CrawlerRunConfig(bypass_cache=True)
+        raw_config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
         raw_result = await crawler.arun(url=raw_html_url, config=raw_config)
 
         if not raw_result.success:
@@ -4152,7 +4152,7 @@ prune_filter = PruningContentFilter(
 For intelligent content filtering and high-quality markdown generation, you can use the **LLMContentFilter**. This filter leverages LLMs to generate relevant markdown while preserving the original content's meaning and structure:
 
 ```python
-from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig, LLMConfig
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig, LLMConfig, DefaultMarkdownGenerator
 from crawl4ai.content_filter_strategy import LLMContentFilter
 
 async def main():
@@ -4175,8 +4175,13 @@ async def main():
         verbose=True
     )
 
+    md_generator = DefaultMarkdownGenerator(
+        content_filter=filter,
+        options={"ignore_links": True}
+    )
+
     config = CrawlerRunConfig(
-        content_filter=filter
+        markdown_generator=md_generator
     )
 
     async with AsyncWebCrawler() as crawler:
@@ -5428,29 +5433,38 @@ Sometimes you need a visual record of a page or a PDF “printout.” Crawl4AI c
 ```python
 import os, asyncio
 from base64 import b64decode
-from crawl4ai import AsyncWebCrawler, CacheMode
+from crawl4ai import AsyncWebCrawler, CacheMode, CrawlerRunConfig
 
 async def main():
+    run_config = CrawlerRunConfig(
+        cache_mode=CacheMode.BYPASS,
+        screenshot=True,
+        pdf=True
+    )
+
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(
             url="https://en.wikipedia.org/wiki/List_of_common_misconceptions",
-            cache_mode=CacheMode.BYPASS,
-            pdf=True,
-            screenshot=True
+            config=run_config
         )
-        
         if result.success:
-            # Save screenshot
+            print(f"Screenshot data present: {result.screenshot is not None}")
+            print(f"PDF data present: {result.pdf is not None}")
+
             if result.screenshot:
+                print(f"[OK] Screenshot captured, size: {len(result.screenshot)} bytes")
                 with open("wikipedia_screenshot.png", "wb") as f:
                     f.write(b64decode(result.screenshot))
-            
-            # Save PDF
+            else:
+                print("[WARN] Screenshot data is None.")
+
             if result.pdf:
+                print(f"[OK] PDF captured, size: {len(result.pdf)} bytes")
                 with open("wikipedia_page.pdf", "wb") as f:
                     f.write(result.pdf)
-            
-            print("[OK] PDF & screenshot captured.")
+            else:
+                print("[WARN] PDF data is None.")
+
         else:
             print("[ERROR]", result.error_message)
 

deploy/docker/job.py

@@ -36,6 +36,7 @@ class LlmJobPayload(BaseModel):
     q:      str
     schema: Optional[str] = None
     cache:  bool = False
+    provider: Optional[str] = None
 
 
 class CrawlJobPayload(BaseModel):
@@ -61,6 +62,7 @@ async def llm_job_enqueue(
         schema=payload.schema,
         cache=payload.cache,
         config=_config,
+        provider=payload.provider,
     )
 
 

deploy/docker/schemas.py

@@ -12,10 +12,10 @@ class CrawlRequest(BaseModel):
 class MarkdownRequest(BaseModel):
     """Request body for the /md endpoint."""
     url: str                    = Field(...,  description="Absolute http/https URL to fetch")
-    f:   FilterType             = Field(FilterType.FIT,
-                                        description="Content‑filter strategy: FIT, RAW, BM25, or LLM")
+    f:   FilterType             = Field(FilterType.FIT, description="Content‑filter strategy: fit, raw, bm25, or llm")
     q:   Optional[str] = Field(None,  description="Query string used by BM25/LLM filters")
     c:   Optional[str] = Field("0",   description="Cache‑bust / revision counter")
+    provider: Optional[str] = Field(None, description="LLM provider override (e.g., 'anthropic/claude-3-opus')")
 
 
 class RawCode(BaseModel):

deploy/docker/server.py

@@ -241,7 +241,7 @@ async def get_markdown(
         raise HTTPException(
             400, "URL must be absolute and start with http/https")
     markdown = await handle_markdown_request(
-        body.url, body.f, body.q, body.c, config
+        body.url, body.f, body.q, body.c, config, body.provider
     )
     return JSONResponse({
         "url": body.url,

deploy/docker/static/playground/index.html

@@ -671,6 +671,16 @@
                         method: 'GET',
                         headers: { 'Accept': 'application/json' }
                     });
+                    responseData = await response.json();
+                    const time = Math.round(performance.now() - startTime);
+                    if (!response.ok) {
+                        updateStatus('error', time);
+                        throw new Error(responseData.error || 'Request failed');
+                    }
+                    updateStatus('success', time);
+                    document.querySelector('#response-content code').textContent = JSON.stringify(responseData, null, 2);
+                    document.querySelector('#response-content code').className = 'json hljs';
+                    forceHighlightElement(document.querySelector('#response-content code'));
                 } else if (endpoint === 'crawl_stream') {
                     // Stream processing
                     response = await fetch(api, {

deploy/docker/utils.py

@@ -1,6 +1,7 @@
 import dns.resolver
 import logging
 import yaml
+import os
 from datetime import datetime
 from enum import Enum
 from pathlib import Path
@@ -19,10 +20,24 @@ class FilterType(str, Enum):
     LLM = "llm"
 
 def load_config() -> Dict:
-    """Load and return application configuration."""
+    """Load and return application configuration with environment variable overrides."""
     config_path = Path(__file__).parent / "config.yml"
     with open(config_path, "r") as config_file:
-        return yaml.safe_load(config_file)
+        config = yaml.safe_load(config_file)
+    
+    # Override LLM provider from environment if set
+    llm_provider = os.environ.get("LLM_PROVIDER")
+    if llm_provider:
+        config["llm"]["provider"] = llm_provider
+        logging.info(f"LLM provider overridden from environment: {llm_provider}")
+    
+    # Also support direct API key from environment if the provider-specific key isn't set
+    llm_api_key = os.environ.get("LLM_API_KEY")
+    if llm_api_key and "api_key" not in config["llm"]:
+        config["llm"]["api_key"] = llm_api_key
+        logging.info("LLM API key loaded from LLM_API_KEY environment variable")
+    
+    return config
 
 def setup_logging(config: Dict) -> None:
     """Configure application logging."""
@@ -56,6 +71,52 @@ def decode_redis_hash(hash_data: Dict[bytes, bytes]) -> Dict[str, str]:
 
 
 
+def get_llm_api_key(config: Dict, provider: Optional[str] = None) -> str:
+    """Get the appropriate API key based on the LLM provider.
+    
+    Args:
+        config: The application configuration dictionary
+        provider: Optional provider override (e.g., "openai/gpt-4")
+    
+    Returns:
+        The API key for the provider, or empty string if not found
+    """
+        
+    # Use provided provider or fall back to config
+    if not provider:
+        provider = config["llm"]["provider"]
+    
+    # Check if direct API key is configured
+    if "api_key" in config["llm"]:
+        return config["llm"]["api_key"]
+    
+    # Fall back to the configured api_key_env if no match
+    return os.environ.get(config["llm"].get("api_key_env", ""), "")
+
+
+def validate_llm_provider(config: Dict, provider: Optional[str] = None) -> tuple[bool, str]:
+    """Validate that the LLM provider has an associated API key.
+    
+    Args:
+        config: The application configuration dictionary
+        provider: Optional provider override (e.g., "openai/gpt-4")
+    
+    Returns:
+        Tuple of (is_valid, error_message)
+    """
+    # Use provided provider or fall back to config
+    if not provider:
+        provider = config["llm"]["provider"]
+    
+    # Get the API key for this provider
+    api_key = get_llm_api_key(config, provider)
+    
+    if not api_key:
+        return False, f"No API key found for provider '{provider}'. Please set the appropriate environment variable."
+    
+    return True, ""
+
+
 def verify_email_domain(email: str) -> bool:
     try:
         domain = email.split('@')[1]

docker-compose.yml

@@ -14,6 +14,7 @@ x-base-config: &base-config
     - TOGETHER_API_KEY=${TOGETHER_API_KEY:-}
     - MISTRAL_API_KEY=${MISTRAL_API_KEY:-}
     - GEMINI_API_TOKEN=${GEMINI_API_TOKEN:-}
+    - LLM_PROVIDER=${LLM_PROVIDER:-}  # Optional: Override default provider (e.g., "anthropic/claude-3-opus")
   volumes:
     - /dev/shm:/dev/shm  # Chromium performance
   deploy:

docs/blog/release-v0.7.0.md

@@ -0,0 +1,343 @@
+# 🚀 Crawl4AI v0.7.0: The Adaptive Intelligence Update
+
+*January 28, 2025 • 10 min read*
+
+---
+
+Today I'm releasing Crawl4AI v0.7.0—the Adaptive Intelligence Update. This release introduces fundamental improvements in how Crawl4AI handles modern web complexity through adaptive learning, intelligent content discovery, and advanced extraction capabilities.
+
+## 🎯 What's New at a Glance
+
+- **Adaptive Crawling**: Your crawler now learns and adapts to website patterns
+- **Virtual Scroll Support**: Complete content extraction from infinite scroll pages
+- **Link Preview with Intelligent Scoring**: Intelligent link analysis and prioritization
+- **Async URL Seeder**: Discover thousands of URLs in seconds with intelligent filtering
+- **Performance Optimizations**: Significant speed and memory improvements
+
+## 🧠 Adaptive Crawling: Intelligence Through Pattern Learning
+
+**The Problem:** Websites change. Class names shift. IDs disappear. Your carefully crafted selectors break at 3 AM, and you wake up to empty datasets and angry stakeholders.
+
+**My Solution:** I implemented an adaptive learning system that observes patterns, builds confidence scores, and adjusts extraction strategies on the fly. It's like having a junior developer who gets better at their job with every page they scrape.
+
+### Technical Deep-Dive
+
+The Adaptive Crawler maintains a persistent state for each domain, tracking:
+- Pattern success rates
+- Selector stability over time  
+- Content structure variations
+- Extraction confidence scores
+
+```python
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+import asyncio
+
+async def main():
+    
+    # Configure adaptive crawler
+    config = AdaptiveConfig(
+        strategy="statistical",  # or "embedding" for semantic understanding
+        max_pages=10,
+        confidence_threshold=0.7,  # Stop at 70% confidence
+        top_k_links=3,  # Follow top 3 links per page
+        min_gain_threshold=0.05  # Need 5% information gain to continue
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        print("Starting adaptive crawl about Python decorators...")
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/glossary.html",
+            query="python decorators functions wrapping"
+        )
+        
+        print(f"\n✅ Crawling Complete!")
+        print(f"• Confidence Level: {adaptive.confidence:.0%}")
+        print(f"• Pages Crawled: {len(result.crawled_urls)}")
+        print(f"• Knowledge Base: {len(adaptive.state.knowledge_base)} documents")
+        
+        # Get most relevant content
+        relevant = adaptive.get_relevant_content(top_k=3)
+        print(f"\nMost Relevant Pages:")
+        for i, page in enumerate(relevant, 1):
+            print(f"{i}. {page['url']} (relevance: {page['score']:.2%})")
+
+asyncio.run(main())
+```
+
+**Expected Real-World Impact:**
+- **News Aggregation**: Maintain 95%+ extraction accuracy even as news sites update their templates
+- **E-commerce Monitoring**: Track product changes across hundreds of stores without constant maintenance
+- **Research Data Collection**: Build robust academic datasets that survive website redesigns
+- **Reduced Maintenance**: Cut selector update time by 80% for frequently-changing sites
+
+## 🌊 Virtual Scroll: Complete Content Capture
+
+**The Problem:** Modern web apps only render what's visible. Scroll down, new content appears, old content vanishes into the void. Traditional crawlers capture that first viewport and miss 90% of the content. It's like reading only the first page of every book.
+
+**My Solution:** I built Virtual Scroll support that mimics human browsing behavior, capturing content as it loads and preserving it before the browser's garbage collector strikes.
+
+### Implementation Details
+
+```python
+from crawl4ai import VirtualScrollConfig
+
+# For social media feeds (Twitter/X style)
+twitter_config = VirtualScrollConfig(
+    container_selector="[data-testid='primaryColumn']",
+    scroll_count=20,                    # Number of scrolls
+    scroll_by="container_height",       # Smart scrolling by container size
+    wait_after_scroll=1.0              # Let content load
+)
+
+# For e-commerce product grids (Instagram style)
+grid_config = VirtualScrollConfig(
+    container_selector="main .product-grid",
+    scroll_count=30,
+    scroll_by=800,                     # Fixed pixel scrolling
+    wait_after_scroll=1.5              # Images need time
+)
+
+# For news feeds with lazy loading
+news_config = VirtualScrollConfig(
+    container_selector=".article-feed",
+    scroll_count=50,
+    scroll_by="page_height",           # Viewport-based scrolling
+    wait_after_scroll=0.5              # Wait for content to load
+)
+
+# Use it in your crawl
+async with AsyncWebCrawler() as crawler:
+    result = await crawler.arun(
+        "https://twitter.com/trending",
+        config=CrawlerRunConfig(
+            virtual_scroll_config=twitter_config,
+            # Combine with other features
+            extraction_strategy=JsonCssExtractionStrategy({
+                "tweets": {
+                    "selector": "[data-testid='tweet']",
+                    "fields": {
+                        "text": {"selector": "[data-testid='tweetText']", "type": "text"},
+                        "likes": {"selector": "[data-testid='like']", "type": "text"}
+                    }
+                }
+            })
+        )
+    )
+    
+    print(f"Captured {len(result.extracted_content['tweets'])} tweets")
+```
+
+**Key Capabilities:**
+- **DOM Recycling Awareness**: Detects and handles virtual DOM element recycling
+- **Smart Scroll Physics**: Three modes - container height, page height, or fixed pixels
+- **Content Preservation**: Captures content before it's destroyed
+- **Intelligent Stopping**: Stops when no new content appears
+- **Memory Efficient**: Streams content instead of holding everything in memory
+
+**Expected Real-World Impact:**
+- **Social Media Analysis**: Capture entire Twitter threads with hundreds of replies, not just top 10
+- **E-commerce Scraping**: Extract 500+ products from infinite scroll catalogs vs. 20-50 with traditional methods  
+- **News Aggregation**: Get all articles from modern news sites, not just above-the-fold content
+- **Research Applications**: Complete data extraction from academic databases using virtual pagination
+
+## 🔗 Link Preview: Intelligent Link Analysis and Scoring
+
+**The Problem:** You crawl a page and get 200 links. Which ones matter? Which lead to the content you actually want? Traditional crawlers force you to follow everything or build complex filters.
+
+**My Solution:** I implemented a three-layer scoring system that analyzes links like a human would—considering their position, context, and relevance to your goals.
+
+### Intelligent Link Analysis and Scoring
+
+```python
+import asyncio
+from crawl4ai import CrawlerRunConfig, CacheMode, AsyncWebCrawler
+from crawl4ai.adaptive_crawler import LinkPreviewConfig
+
+async def main():
+    # Configure intelligent link analysis
+    link_config = LinkPreviewConfig(
+        include_internal=True,
+        include_external=False,
+        max_links=10,
+        concurrency=5,
+        query="python tutorial",  # For contextual scoring
+        score_threshold=0.3,
+        verbose=True
+    )
+    # Use in your crawl
+    async with AsyncWebCrawler() as crawler:
+        result = await crawler.arun(
+            "https://www.geeksforgeeks.org/",
+            config=CrawlerRunConfig(
+                link_preview_config=link_config,
+                score_links=True,  # Enable intrinsic scoring
+                cache_mode=CacheMode.BYPASS
+            )
+        )
+
+        # Access scored and sorted links
+        if result.success and result.links:
+            for link in result.links.get("internal", []):
+                text = link.get('text', 'No text')[:40]
+                print(
+                    text,
+                    f"{link.get('intrinsic_score', 0):.1f}/10" if link.get('intrinsic_score') is not None else "0.0/10",
+                    f"{link.get('contextual_score', 0):.2f}/1" if link.get('contextual_score') is not None else "0.00/1",
+                    f"{link.get('total_score', 0):.3f}" if link.get('total_score') is not None else "0.000"
+                )
+
+asyncio.run(main())
+```
+
+**Scoring Components:**
+
+1. **Intrinsic Score**: Based on link quality indicators
+   - Position on page (navigation, content, footer)
+   - Link attributes (rel, title, class names)
+   - Anchor text quality and length
+   - URL structure and depth
+
+2. **Contextual Score**: Relevance to your query using BM25 algorithm
+   - Keyword matching in link text and title
+   - Meta description analysis
+   - Content preview scoring
+
+3. **Total Score**: Combined score for final ranking
+
+**Expected Real-World Impact:**
+- **Research Efficiency**: Find relevant papers 10x faster by following only high-score links
+- **Competitive Analysis**: Automatically identify important pages on competitor sites
+- **Content Discovery**: Build topic-focused crawlers that stay on track
+- **SEO Audits**: Identify and prioritize high-value internal linking opportunities
+
+## 🎣 Async URL Seeder: Automated URL Discovery at Scale
+
+**The Problem:** You want to crawl an entire domain but only have the homepage. Or worse, you want specific content types across thousands of pages. Manual URL discovery? That's a job for machines, not humans.
+
+**My Solution:** I built Async URL Seeder—a turbocharged URL discovery engine that combines multiple sources with intelligent filtering and relevance scoring.
+
+### Technical Architecture
+
+```python
+import asyncio
+from crawl4ai import AsyncUrlSeeder, SeedingConfig
+
+async def main():
+    async with AsyncUrlSeeder() as seeder:
+        # Discover Python tutorial URLs
+        config = SeedingConfig(
+            source="sitemap",  # Use sitemap
+            pattern="*python*",  # URL pattern filter
+            extract_head=True,  # Get metadata
+            query="python tutorial",  # For relevance scoring
+            scoring_method="bm25",
+            score_threshold=0.2,
+            max_urls=10
+        )
+        
+        print("Discovering Python async tutorial URLs...")
+        urls = await seeder.urls("https://www.geeksforgeeks.org/", config)
+        
+        print(f"\n✅ Found {len(urls)} relevant URLs:")
+        for i, url_info in enumerate(urls[:5], 1):
+            print(f"\n{i}. {url_info['url']}")
+            if url_info.get('relevance_score'):
+                print(f"   Relevance: {url_info['relevance_score']:.3f}")
+            if url_info.get('head_data', {}).get('title'):
+                print(f"   Title: {url_info['head_data']['title'][:60]}...")
+
+asyncio.run(main())
+```
+
+**Discovery Methods:**
+- **Sitemap Mining**: Parses robots.txt and all linked sitemaps
+- **Common Crawl**: Queries the Common Crawl index for historical URLs
+- **Intelligent Crawling**: Follows links with smart depth control
+- **Pattern Analysis**: Learns URL structures and generates variations
+
+**Expected Real-World Impact:**
+- **Migration Projects**: Discover 10,000+ URLs from legacy sites in under 60 seconds
+- **Market Research**: Map entire competitor ecosystems automatically
+- **Academic Research**: Build comprehensive datasets without manual URL collection
+- **SEO Audits**: Find every indexable page with content scoring
+- **Content Archival**: Ensure no content is left behind during site migrations
+
+## ⚡ Performance Optimizations
+
+This release includes significant performance improvements through optimized resource handling, better concurrency management, and reduced memory footprint.
+
+### What We Optimized
+
+```python
+# Optimized crawling with v0.7.0 improvements
+results = []
+for url in urls:
+    result = await crawler.arun(
+        url,
+        config=CrawlerRunConfig(
+            # Performance optimizations
+            wait_until="domcontentloaded",  # Faster than networkidle
+            cache_mode=CacheMode.ENABLED    # Enable caching
+        )
+    )
+    results.append(result)
+```
+
+**Performance Gains:**
+- **Startup Time**: 70% faster browser initialization
+- **Page Loading**: 40% reduction with smart resource blocking
+- **Extraction**: 3x faster with compiled CSS selectors
+- **Memory Usage**: 60% reduction with streaming processing
+- **Concurrent Crawls**: Handle 5x more parallel requests
+
+
+## 🔧 Important Changes
+
+### Breaking Changes
+- `link_extractor` renamed to `link_preview` (better reflects functionality)
+- Minimum Python version now 3.9
+- `CrawlerConfig` split into `CrawlerRunConfig` and `BrowserConfig`
+
+### Migration Guide
+```python
+# Old (v0.6.x)
+from crawl4ai import CrawlerConfig
+config = CrawlerConfig(timeout=30000)
+
+# New (v0.7.0)
+from crawl4ai import CrawlerRunConfig, BrowserConfig
+browser_config = BrowserConfig(timeout=30000)
+run_config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
+```
+
+## 🤖 Coming Soon: Intelligent Web Automation
+
+I'm currently working on bringing advanced automation capabilities to Crawl4AI. This includes:
+
+- **Crawl Agents**: Autonomous crawlers that understand your goals and adapt their strategies
+- **Auto JS Generation**: Automatic JavaScript code generation for complex interactions
+- **Smart Form Handling**: Intelligent form detection and filling
+- **Context-Aware Actions**: Crawlers that understand page context and make decisions
+
+These features are under active development and will revolutionize how we approach web automation. Stay tuned!
+
+## 🚀 Get Started
+
+```bash
+pip install crawl4ai==0.7.0
+```
+
+Check out the [updated documentation](https://docs.crawl4ai.com).
+
+Questions? Issues? I'm always listening:
+- GitHub: [github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+- Discord: [discord.gg/crawl4ai](https://discord.gg/jP8KfhDhyN)
+- Twitter: [@unclecode](https://x.com/unclecode)
+
+Happy crawling! 🕷️
+
+---
+
+*P.S. If you're using Crawl4AI in production, I'd love to hear about it. Your use cases inspire the next features.*

docs/blog/release-v0.7.1.md

@@ -0,0 +1,43 @@
+# 🛠️ Crawl4AI v0.7.1: Minor Cleanup Update
+
+*July 17, 2025 • 2 min read*
+
+---
+
+A small maintenance release that removes unused code and improves documentation.
+
+## 🎯 What's Changed
+
+- **Removed unused StealthConfig** from `crawl4ai/browser_manager.py`
+- **Updated documentation** with better examples and parameter explanations
+- **Fixed virtual scroll configuration** examples in docs
+
+## 🧹 Code Cleanup
+
+Removed unused `StealthConfig` import and configuration that wasn't being used anywhere in the codebase. The project uses its own custom stealth implementation through JavaScript injection instead.
+
+```python
+# Removed unused code:
+from playwright_stealth import StealthConfig
+stealth_config = StealthConfig(...)  # This was never used
+```
+
+## 📖 Documentation Updates
+
+- Fixed adaptive crawling parameter examples
+- Updated session management documentation
+- Corrected virtual scroll configuration examples
+
+## 🚀 Installation
+
+```bash
+pip install crawl4ai==0.7.1
+```
+
+No breaking changes - upgrade directly from v0.7.0.
+
+---
+
+Questions? Issues? 
+- GitHub: [github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+- Discord: [discord.gg/crawl4ai](https://discord.gg/jP8KfhDhyN)

docs/examples/adaptive_crawling/README.md

@@ -0,0 +1,85 @@
+# Adaptive Crawling Examples
+
+This directory contains examples demonstrating various aspects of Crawl4AI's Adaptive Crawling feature.
+
+## Examples Overview
+
+### 1. `basic_usage.py`
+- Simple introduction to adaptive crawling
+- Uses default statistical strategy
+- Shows how to get crawl statistics and relevant content
+
+### 2. `embedding_strategy.py` ⭐ NEW
+- Demonstrates the embedding-based strategy for semantic understanding
+- Shows query expansion and irrelevance detection
+- Includes configuration for both local and API-based embeddings
+
+### 3. `embedding_vs_statistical.py` ⭐ NEW
+- Direct comparison between statistical and embedding strategies
+- Helps you choose the right strategy for your use case
+- Shows performance and accuracy trade-offs
+
+### 4. `embedding_configuration.py` ⭐ NEW
+- Advanced configuration options for embedding strategy
+- Parameter tuning guide for different scenarios
+- Examples for research, exploration, and quality-focused crawling
+
+### 5. `advanced_configuration.py`
+- Shows various configuration options for both strategies
+- Demonstrates threshold tuning and performance optimization
+
+### 6. `custom_strategies.py`
+- How to implement your own crawling strategy
+- Extends the base CrawlStrategy class
+- Advanced use case for specialized requirements
+
+### 7. `export_import_kb.py`
+- Export crawled knowledge base to JSONL
+- Import and continue crawling from saved state
+- Useful for building persistent knowledge bases
+
+## Quick Start
+
+For your first adaptive crawling experience, run:
+
+```bash
+python basic_usage.py
+```
+
+To try the new embedding strategy with semantic understanding:
+
+```bash
+python embedding_strategy.py
+```
+
+To compare strategies and see which works best for your use case:
+
+```bash
+python embedding_vs_statistical.py
+```
+
+## Strategy Selection Guide
+
+### Use Statistical Strategy (Default) When:
+- Working with technical documentation
+- Queries contain specific terms or code
+- Speed is critical
+- No API access available
+
+### Use Embedding Strategy When:
+- Queries are conceptual or ambiguous
+- Need semantic understanding beyond exact matches
+- Want to detect irrelevant content
+- Working with diverse content sources
+
+## Requirements
+
+- Crawl4AI installed
+- For embedding strategy with local models: `sentence-transformers`
+- For embedding strategy with OpenAI: Set `OPENAI_API_KEY` environment variable
+
+## Learn More
+
+- [Adaptive Crawling Documentation](https://docs.crawl4ai.com/core/adaptive-crawling/)
+- [Mathematical Framework](https://github.com/unclecode/crawl4ai/blob/main/PROGRESSIVE_CRAWLING.md)
+- [Blog: The Adaptive Crawling Revolution](https://docs.crawl4ai.com/blog/adaptive-crawling-revolution/)

docs/examples/adaptive_crawling/advanced_configuration.py

@@ -0,0 +1,207 @@
+"""
+Advanced Adaptive Crawling Configuration
+
+This example demonstrates all configuration options available for adaptive crawling,
+including threshold tuning, persistence, and custom parameters.
+"""
+
+import asyncio
+from pathlib import Path
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+
+async def main():
+    """Demonstrate advanced configuration options"""
+    
+    # Example 1: Custom thresholds for different use cases
+    print("="*60)
+    print("EXAMPLE 1: Custom Confidence Thresholds")
+    print("="*60)
+    
+    # High-precision configuration (exhaustive crawling)
+    high_precision_config = AdaptiveConfig(
+        confidence_threshold=0.9,      # Very high confidence required
+        max_pages=50,                  # Allow more pages
+        top_k_links=5,                 # Follow more links per page
+        min_gain_threshold=0.02        # Lower threshold to continue
+    )
+    
+    # Balanced configuration (default use case)
+    balanced_config = AdaptiveConfig(
+        confidence_threshold=0.7,      # Moderate confidence
+        max_pages=20,                  # Reasonable limit
+        top_k_links=3,                 # Moderate branching
+        min_gain_threshold=0.05        # Standard gain threshold
+    )
+    
+    # Quick exploration configuration
+    quick_config = AdaptiveConfig(
+        confidence_threshold=0.5,      # Lower confidence acceptable
+        max_pages=10,                  # Strict limit
+        top_k_links=2,                 # Minimal branching
+        min_gain_threshold=0.1         # High gain required
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        # Test different configurations
+        for config_name, config in [
+            ("High Precision", high_precision_config),
+            ("Balanced", balanced_config),
+            ("Quick Exploration", quick_config)
+        ]:
+            print(f"\nTesting {config_name} configuration...")
+            adaptive = AdaptiveCrawler(crawler, config=config)
+            
+            result = await adaptive.digest(
+                start_url="https://httpbin.org",
+                query="http headers authentication"
+            )
+            
+            print(f"  - Pages crawled: {len(result.crawled_urls)}")
+            print(f"  - Confidence achieved: {adaptive.confidence:.2%}")
+            print(f"  - Coverage score: {adaptive.coverage_stats['coverage']:.2f}")
+    
+    # Example 2: Persistence and state management
+    print("\n" + "="*60)
+    print("EXAMPLE 2: State Persistence")
+    print("="*60)
+    
+    state_file = "crawl_state_demo.json"
+    
+    # Configuration with persistence
+    persistent_config = AdaptiveConfig(
+        confidence_threshold=0.8,
+        max_pages=30,
+        save_state=True,              # Enable auto-save
+        state_path=state_file         # Specify save location
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        # First crawl - will be interrupted
+        print("\nStarting initial crawl (will interrupt after 5 pages)...")
+        
+        interrupt_config = AdaptiveConfig(
+            confidence_threshold=0.8,
+            max_pages=5,              # Artificially low to simulate interruption
+            save_state=True,
+            state_path=state_file
+        )
+        
+        adaptive = AdaptiveCrawler(crawler, config=interrupt_config)
+        result1 = await adaptive.digest(
+            start_url="https://docs.python.org/3/",
+            query="exception handling try except finally"
+        )
+        
+        print(f"First crawl completed: {len(result1.crawled_urls)} pages")
+        print(f"Confidence reached: {adaptive.confidence:.2%}")
+        
+        # Resume crawl with higher page limit
+        print("\nResuming crawl from saved state...")
+        
+        resume_config = AdaptiveConfig(
+            confidence_threshold=0.8,
+            max_pages=20,             # Increase limit
+            save_state=True,
+            state_path=state_file
+        )
+        
+        adaptive2 = AdaptiveCrawler(crawler, config=resume_config)
+        result2 = await adaptive2.digest(
+            start_url="https://docs.python.org/3/",
+            query="exception handling try except finally",
+            resume_from=state_file
+        )
+        
+        print(f"Resumed crawl completed: {len(result2.crawled_urls)} total pages")
+        print(f"Final confidence: {adaptive2.confidence:.2%}")
+        
+        # Clean up
+        Path(state_file).unlink(missing_ok=True)
+    
+    # Example 3: Link selection strategies
+    print("\n" + "="*60)
+    print("EXAMPLE 3: Link Selection Strategies")
+    print("="*60)
+    
+    # Conservative link following
+    conservative_config = AdaptiveConfig(
+        confidence_threshold=0.7,
+        max_pages=15,
+        top_k_links=1,                # Only follow best link
+        min_gain_threshold=0.15       # High threshold
+    )
+    
+    # Aggressive link following
+    aggressive_config = AdaptiveConfig(
+        confidence_threshold=0.7,
+        max_pages=15,
+        top_k_links=10,               # Follow many links
+        min_gain_threshold=0.01       # Very low threshold
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        for strategy_name, config in [
+            ("Conservative", conservative_config),
+            ("Aggressive", aggressive_config)
+        ]:
+            print(f"\n{strategy_name} link selection:")
+            adaptive = AdaptiveCrawler(crawler, config=config)
+            
+            result = await adaptive.digest(
+                start_url="https://httpbin.org",
+                query="api endpoints"
+            )
+            
+            # Analyze crawl pattern
+            print(f"  - Total pages: {len(result.crawled_urls)}")
+            print(f"  - Unique domains: {len(set(url.split('/')[2] for url in result.crawled_urls))}")
+            print(f"  - Max depth reached: {max(url.count('/') for url in result.crawled_urls) - 2}")
+            
+            # Show saturation trend
+            if hasattr(result, 'new_terms_history') and result.new_terms_history:
+                print(f"  - New terms discovered: {result.new_terms_history[:5]}...")
+                print(f"  - Saturation trend: {'decreasing' if result.new_terms_history[-1] < result.new_terms_history[0] else 'increasing'}")
+    
+    # Example 4: Monitoring crawl progress
+    print("\n" + "="*60)
+    print("EXAMPLE 4: Progress Monitoring")
+    print("="*60)
+    
+    # Configuration with detailed monitoring
+    monitor_config = AdaptiveConfig(
+        confidence_threshold=0.75,
+        max_pages=10,
+        top_k_links=3
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config=monitor_config)
+        
+        # Start crawl
+        print("\nMonitoring crawl progress...")
+        result = await adaptive.digest(
+            start_url="https://httpbin.org",
+            query="http methods headers"
+        )
+        
+        # Detailed statistics
+        print("\nDetailed crawl analysis:")
+        adaptive.print_stats(detailed=True)
+        
+        # Export for analysis
+        print("\nExporting knowledge base for external analysis...")
+        adaptive.export_knowledge_base("knowledge_export_demo.jsonl")
+        print("Knowledge base exported to: knowledge_export_demo.jsonl")
+        
+        # Show sample of exported data
+        with open("knowledge_export_demo.jsonl", 'r') as f:
+            first_line = f.readline()
+            print(f"Sample export: {first_line[:100]}...")
+        
+        # Clean up
+        Path("knowledge_export_demo.jsonl").unlink(missing_ok=True)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/basic_usage.py

@@ -0,0 +1,76 @@
+"""
+Basic Adaptive Crawling Example
+
+This example demonstrates the simplest use case of adaptive crawling:
+finding information about a specific topic and knowing when to stop.
+"""
+
+import asyncio
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler
+
+
+async def main():
+    """Basic adaptive crawling example"""
+    
+    # Initialize the crawler
+    async with AsyncWebCrawler(verbose=True) as crawler:
+        # Create an adaptive crawler with default settings (statistical strategy)
+        adaptive = AdaptiveCrawler(crawler)
+        
+        # Note: You can also use embedding strategy for semantic understanding:
+        # from crawl4ai import AdaptiveConfig
+        # config = AdaptiveConfig(strategy="embedding")
+        # adaptive = AdaptiveCrawler(crawler, config)
+        
+        # Start adaptive crawling
+        print("Starting adaptive crawl for Python async programming information...")
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/library/asyncio.html",
+            query="async await context managers coroutines"
+        )
+        
+        # Display crawl statistics
+        print("\n" + "="*50)
+        print("CRAWL STATISTICS")
+        print("="*50)
+        adaptive.print_stats(detailed=False)
+        
+        # Get the most relevant content found
+        print("\n" + "="*50)
+        print("MOST RELEVANT PAGES")
+        print("="*50)
+        
+        relevant_pages = adaptive.get_relevant_content(top_k=5)
+        for i, page in enumerate(relevant_pages, 1):
+            print(f"\n{i}. {page['url']}")
+            print(f"   Relevance Score: {page['score']:.2%}")
+            
+            # Show a snippet of the content
+            content = page['content'] or ""
+            if content:
+                snippet = content[:200].replace('\n', ' ')
+                if len(content) > 200:
+                    snippet += "..."
+                print(f"   Preview: {snippet}")
+        
+        # Show final confidence
+        print(f"\n{'='*50}")
+        print(f"Final Confidence: {adaptive.confidence:.2%}")
+        print(f"Total Pages Crawled: {len(result.crawled_urls)}")
+        print(f"Knowledge Base Size: {len(adaptive.state.knowledge_base)} documents")
+        
+        # Example: Check if we can answer specific questions
+        print(f"\n{'='*50}")
+        print("INFORMATION SUFFICIENCY CHECK")
+        print(f"{'='*50}")
+        
+        if adaptive.confidence >= 0.8:
+            print("✓ High confidence - can answer detailed questions about async Python")
+        elif adaptive.confidence >= 0.6:
+            print("~ Moderate confidence - can answer basic questions") 
+        else:
+            print("✗ Low confidence - need more information")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/custom_strategies.py

@@ -0,0 +1,373 @@
+"""
+Custom Adaptive Crawling Strategies
+
+This example demonstrates how to implement custom scoring strategies
+for domain-specific crawling needs.
+"""
+
+import asyncio
+import re
+from typing import List, Dict, Set
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+from crawl4ai.adaptive_crawler import CrawlState, Link
+import math
+
+
+class APIDocumentationStrategy:
+    """
+    Custom strategy optimized for API documentation crawling.
+    Prioritizes endpoint references, code examples, and parameter descriptions.
+    """
+    
+    def __init__(self):
+        # Keywords that indicate high-value API documentation
+        self.api_keywords = {
+            'endpoint', 'request', 'response', 'parameter', 'authentication',
+            'header', 'body', 'query', 'path', 'method', 'get', 'post', 'put',
+            'delete', 'patch', 'status', 'code', 'example', 'curl', 'python'
+        }
+        
+        # URL patterns that typically contain API documentation
+        self.valuable_patterns = [
+            r'/api/',
+            r'/reference/',
+            r'/endpoints?/',
+            r'/methods?/',
+            r'/resources?/'
+        ]
+        
+        # Patterns to avoid
+        self.avoid_patterns = [
+            r'/blog/',
+            r'/news/',
+            r'/about/',
+            r'/contact/',
+            r'/legal/'
+        ]
+    
+    def score_link(self, link: Link, query: str, state: CrawlState) -> float:
+        """Custom link scoring for API documentation"""
+        score = 1.0
+        url = link.href.lower()
+        
+        # Boost API-related URLs
+        for pattern in self.valuable_patterns:
+            if re.search(pattern, url):
+                score *= 2.0
+                break
+        
+        # Reduce score for non-API content
+        for pattern in self.avoid_patterns:
+            if re.search(pattern, url):
+                score *= 0.1
+                break
+        
+        # Boost if preview contains API keywords
+        if link.text:
+            preview_lower = link.text.lower()
+            keyword_count = sum(1 for kw in self.api_keywords if kw in preview_lower)
+            score *= (1 + keyword_count * 0.2)
+        
+        # Prioritize shallow URLs (likely overview pages)
+        depth = url.count('/') - 2  # Subtract protocol slashes
+        if depth <= 3:
+            score *= 1.5
+        elif depth > 6:
+            score *= 0.5
+        
+        return score
+    
+    def calculate_api_coverage(self, state: CrawlState, query: str) -> Dict[str, float]:
+        """Calculate specialized coverage metrics for API documentation"""
+        metrics = {
+            'endpoint_coverage': 0.0,
+            'example_coverage': 0.0,
+            'parameter_coverage': 0.0
+        }
+        
+        # Analyze knowledge base for API-specific content
+        endpoint_patterns = [r'GET\s+/', r'POST\s+/', r'PUT\s+/', r'DELETE\s+/']
+        example_patterns = [r'```\w+', r'curl\s+-', r'import\s+requests']
+        param_patterns = [r'param(?:eter)?s?\s*:', r'required\s*:', r'optional\s*:']
+        
+        total_docs = len(state.knowledge_base)
+        if total_docs == 0:
+            return metrics
+        
+        docs_with_endpoints = 0
+        docs_with_examples = 0
+        docs_with_params = 0
+        
+        for doc in state.knowledge_base:
+            content = doc.markdown.raw_markdown if hasattr(doc, 'markdown') else str(doc)
+            
+            # Check for endpoints
+            if any(re.search(pattern, content, re.IGNORECASE) for pattern in endpoint_patterns):
+                docs_with_endpoints += 1
+            
+            # Check for examples
+            if any(re.search(pattern, content, re.IGNORECASE) for pattern in example_patterns):
+                docs_with_examples += 1
+            
+            # Check for parameters
+            if any(re.search(pattern, content, re.IGNORECASE) for pattern in param_patterns):
+                docs_with_params += 1
+        
+        metrics['endpoint_coverage'] = docs_with_endpoints / total_docs
+        metrics['example_coverage'] = docs_with_examples / total_docs
+        metrics['parameter_coverage'] = docs_with_params / total_docs
+        
+        return metrics
+
+
+class ResearchPaperStrategy:
+    """
+    Strategy optimized for crawling research papers and academic content.
+    Prioritizes citations, abstracts, and methodology sections.
+    """
+    
+    def __init__(self):
+        self.academic_keywords = {
+            'abstract', 'introduction', 'methodology', 'results', 'conclusion',
+            'references', 'citation', 'paper', 'study', 'research', 'analysis',
+            'hypothesis', 'experiment', 'findings', 'doi'
+        }
+        
+        self.citation_patterns = [
+            r'\[\d+\]',  # [1] style citations
+            r'\(\w+\s+\d{4}\)',  # (Author 2024) style
+            r'doi:\s*\S+',  # DOI references
+        ]
+    
+    def calculate_academic_relevance(self, content: str, query: str) -> float:
+        """Calculate relevance score for academic content"""
+        score = 0.0
+        content_lower = content.lower()
+        
+        # Check for academic keywords
+        keyword_matches = sum(1 for kw in self.academic_keywords if kw in content_lower)
+        score += keyword_matches * 0.1
+        
+        # Check for citations
+        citation_count = sum(
+            len(re.findall(pattern, content)) 
+            for pattern in self.citation_patterns
+        )
+        score += min(citation_count * 0.05, 1.0)  # Cap at 1.0
+        
+        # Check for query terms in academic context
+        query_terms = query.lower().split()
+        for term in query_terms:
+            # Boost if term appears near academic keywords
+            for keyword in ['abstract', 'conclusion', 'results']:
+                if keyword in content_lower:
+                    section = content_lower[content_lower.find(keyword):content_lower.find(keyword) + 500]
+                    if term in section:
+                        score += 0.2
+        
+        return min(score, 2.0)  # Cap total score
+
+
+async def demo_custom_strategies():
+    """Demonstrate custom strategy usage"""
+    
+    # Example 1: API Documentation Strategy
+    print("="*60)
+    print("EXAMPLE 1: Custom API Documentation Strategy")
+    print("="*60)
+    
+    api_strategy = APIDocumentationStrategy()
+    
+    async with AsyncWebCrawler() as crawler:
+        # Standard adaptive crawler
+        config = AdaptiveConfig(
+            confidence_threshold=0.8,
+            max_pages=15
+        )
+        
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        # Override link scoring with custom strategy
+        original_rank_links = adaptive._rank_links
+        
+        def custom_rank_links(links, query, state):
+            # Apply custom scoring
+            scored_links = []
+            for link in links:
+                base_score = api_strategy.score_link(link, query, state)
+                scored_links.append((link, base_score))
+            
+            # Sort by score
+            scored_links.sort(key=lambda x: x[1], reverse=True)
+            return [link for link, _ in scored_links[:config.top_k_links]]
+        
+        adaptive._rank_links = custom_rank_links
+        
+        # Crawl API documentation
+        print("\nCrawling API documentation with custom strategy...")
+        state = await adaptive.digest(
+            start_url="https://httpbin.org",
+            query="api endpoints authentication headers"
+        )
+        
+        # Calculate custom metrics
+        api_metrics = api_strategy.calculate_api_coverage(state, "api endpoints")
+        
+        print(f"\nResults:")
+        print(f"Pages crawled: {len(state.crawled_urls)}")
+        print(f"Confidence: {adaptive.confidence:.2%}")
+        print(f"\nAPI-Specific Metrics:")
+        print(f"  - Endpoint coverage: {api_metrics['endpoint_coverage']:.2%}")
+        print(f"  - Example coverage: {api_metrics['example_coverage']:.2%}")
+        print(f"  - Parameter coverage: {api_metrics['parameter_coverage']:.2%}")
+    
+    # Example 2: Combined Strategy
+    print("\n" + "="*60)
+    print("EXAMPLE 2: Hybrid Strategy Combining Multiple Approaches")
+    print("="*60)
+    
+    class HybridStrategy:
+        """Combines multiple strategies with weights"""
+        
+        def __init__(self):
+            self.api_strategy = APIDocumentationStrategy()
+            self.research_strategy = ResearchPaperStrategy()
+            self.weights = {
+                'api': 0.7,
+                'research': 0.3
+            }
+        
+        def score_content(self, content: str, query: str) -> float:
+            # Get scores from each strategy
+            api_score = self._calculate_api_score(content, query)
+            research_score = self.research_strategy.calculate_academic_relevance(content, query)
+            
+            # Weighted combination
+            total_score = (
+                api_score * self.weights['api'] +
+                research_score * self.weights['research']
+            )
+            
+            return total_score
+        
+        def _calculate_api_score(self, content: str, query: str) -> float:
+            # Simplified API scoring based on keyword presence
+            content_lower = content.lower()
+            api_keywords = self.api_strategy.api_keywords
+            
+            keyword_count = sum(1 for kw in api_keywords if kw in content_lower)
+            return min(keyword_count * 0.1, 2.0)
+    
+    hybrid_strategy = HybridStrategy()
+    
+    async with AsyncWebCrawler() as crawler:
+        adaptive = AdaptiveCrawler(crawler)
+        
+        # Crawl with hybrid scoring
+        print("\nTesting hybrid strategy on technical documentation...")
+        state = await adaptive.digest(
+            start_url="https://docs.python.org/3/library/asyncio.html",
+            query="async await coroutines api"
+        )
+        
+        # Analyze results with hybrid strategy
+        print(f"\nHybrid Strategy Analysis:")
+        total_score = 0
+        for doc in adaptive.get_relevant_content(top_k=5):
+            content = doc['content'] or ""
+            score = hybrid_strategy.score_content(content, "async await api")
+            total_score += score
+            print(f"  - {doc['url'][:50]}... Score: {score:.2f}")
+        
+        print(f"\nAverage hybrid score: {total_score/5:.2f}")
+
+
+async def demo_performance_optimization():
+    """Demonstrate performance optimization with custom strategies"""
+    
+    print("\n" + "="*60)
+    print("EXAMPLE 3: Performance-Optimized Strategy")
+    print("="*60)
+    
+    class PerformanceOptimizedStrategy:
+        """Strategy that balances thoroughness with speed"""
+        
+        def __init__(self):
+            self.url_cache: Set[str] = set()
+            self.domain_scores: Dict[str, float] = {}
+        
+        def should_crawl_domain(self, url: str) -> bool:
+            """Implement domain-level filtering"""
+            domain = url.split('/')[2] if url.startswith('http') else url
+            
+            # Skip if we've already crawled many pages from this domain
+            domain_count = sum(1 for cached in self.url_cache if domain in cached)
+            if domain_count > 5:
+                return False
+            
+            # Skip low-scoring domains
+            if domain in self.domain_scores and self.domain_scores[domain] < 0.3:
+                return False
+            
+            return True
+        
+        def update_domain_score(self, url: str, relevance: float):
+            """Track domain-level performance"""
+            domain = url.split('/')[2] if url.startswith('http') else url
+            
+            if domain not in self.domain_scores:
+                self.domain_scores[domain] = relevance
+            else:
+                # Moving average
+                self.domain_scores[domain] = (
+                    0.7 * self.domain_scores[domain] + 0.3 * relevance
+                )
+    
+    perf_strategy = PerformanceOptimizedStrategy()
+    
+    async with AsyncWebCrawler() as crawler:
+        config = AdaptiveConfig(
+            confidence_threshold=0.7,
+            max_pages=10,
+            top_k_links=2  # Fewer links for speed
+        )
+        
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        # Track performance
+        import time
+        start_time = time.time()
+        
+        state = await adaptive.digest(
+            start_url="https://httpbin.org",
+            query="http methods headers"
+        )
+        
+        elapsed = time.time() - start_time
+        
+        print(f"\nPerformance Results:")
+        print(f"  - Time elapsed: {elapsed:.2f} seconds")
+        print(f"  - Pages crawled: {len(state.crawled_urls)}")
+        print(f"  - Pages per second: {len(state.crawled_urls)/elapsed:.2f}")
+        print(f"  - Final confidence: {adaptive.confidence:.2%}")
+        print(f"  - Efficiency: {adaptive.confidence/len(state.crawled_urls):.2%} confidence per page")
+
+
+async def main():
+    """Run all demonstrations"""
+    try:
+        await demo_custom_strategies()
+        await demo_performance_optimization()
+        
+        print("\n" + "="*60)
+        print("All custom strategy examples completed!")
+        print("="*60)
+        
+    except Exception as e:
+        print(f"Error: {e}")
+        import traceback
+        traceback.print_exc()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/embedding_configuration.py

@@ -0,0 +1,206 @@
+"""
+Advanced Embedding Configuration Example
+
+This example demonstrates all configuration options available for the
+embedding strategy, including fine-tuning parameters for different use cases.
+"""
+
+import asyncio
+import os
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+
+async def test_configuration(name: str, config: AdaptiveConfig, url: str, query: str):
+    """Test a specific configuration"""
+    print(f"\n{'='*60}")
+    print(f"Configuration: {name}")
+    print(f"{'='*60}")
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        result = await adaptive.digest(start_url=url, query=query)
+        
+        print(f"Pages crawled: {len(result.crawled_urls)}")
+        print(f"Final confidence: {adaptive.confidence:.1%}")
+        print(f"Stopped reason: {result.metrics.get('stopped_reason', 'max_pages')}")
+        
+        if result.metrics.get('is_irrelevant', False):
+            print("⚠️  Query detected as irrelevant!")
+        
+        return result
+
+
+async def main():
+    """Demonstrate various embedding configurations"""
+    
+    print("EMBEDDING STRATEGY CONFIGURATION EXAMPLES")
+    print("=" * 60)
+    
+    # Base URL and query for testing
+    test_url = "https://docs.python.org/3/library/asyncio.html"
+    
+    # 1. Default Configuration
+    config_default = AdaptiveConfig(
+        strategy="embedding",
+        max_pages=10
+    )
+    
+    await test_configuration(
+        "Default Settings",
+        config_default,
+        test_url,
+        "async programming patterns"
+    )
+    
+    # 2. Strict Coverage Requirements
+    config_strict = AdaptiveConfig(
+        strategy="embedding",
+        max_pages=20,
+        
+        # Stricter similarity requirements
+        embedding_k_exp=5.0,  # Default is 3.0, higher = stricter
+        embedding_coverage_radius=0.15,  # Default is 0.2, lower = stricter
+        
+        # Higher validation threshold
+        embedding_validation_min_score=0.6,  # Default is 0.3
+        
+        # More query variations for better coverage
+        n_query_variations=15  # Default is 10
+    )
+    
+    await test_configuration(
+        "Strict Coverage (Research/Academic)",
+        config_strict,
+        test_url,
+        "comprehensive guide async await"
+    )
+    
+    # 3. Fast Exploration
+    config_fast = AdaptiveConfig(
+        strategy="embedding",
+        max_pages=10,
+        top_k_links=5,  # Follow more links per page
+        
+        # Relaxed requirements for faster convergence
+        embedding_k_exp=1.0,  # Lower = more lenient
+        embedding_min_relative_improvement=0.05,  # Stop earlier
+        
+        # Lower quality thresholds
+        embedding_quality_min_confidence=0.5,  # Display lower confidence
+        embedding_quality_max_confidence=0.85,
+        
+        # Fewer query variations for speed
+        n_query_variations=5
+    )
+    
+    await test_configuration(
+        "Fast Exploration (Quick Overview)",
+        config_fast,
+        test_url,
+        "async basics"
+    )
+    
+    # 4. Irrelevance Detection Focus
+    config_irrelevance = AdaptiveConfig(
+        strategy="embedding",
+        max_pages=5,
+        
+        # Aggressive irrelevance detection
+        embedding_min_confidence_threshold=0.2,  # Higher threshold (default 0.1)
+        embedding_k_exp=5.0,  # Strict similarity
+        
+        # Quick stopping for irrelevant content
+        embedding_min_relative_improvement=0.15
+    )
+    
+    await test_configuration(
+        "Irrelevance Detection",
+        config_irrelevance,
+        test_url,
+        "recipe for chocolate cake"  # Irrelevant query
+    )
+    
+    # 5. High-Quality Knowledge Base
+    config_quality = AdaptiveConfig(
+        strategy="embedding",
+        max_pages=30,
+        
+        # Deduplication settings
+        embedding_overlap_threshold=0.75,  # More aggressive deduplication
+        
+        # Quality focus
+        embedding_validation_min_score=0.5,
+        embedding_quality_scale_factor=1.0,  # Linear quality mapping
+        
+        # Balanced parameters
+        embedding_k_exp=3.0,
+        embedding_nearest_weight=0.8,  # Focus on best matches
+        embedding_top_k_weight=0.2
+    )
+    
+    await test_configuration(
+        "High-Quality Knowledge Base",
+        config_quality,
+        test_url,
+        "asyncio advanced patterns best practices"
+    )
+    
+    # 6. Custom Embedding Provider
+    if os.getenv('OPENAI_API_KEY'):
+        config_openai = AdaptiveConfig(
+            strategy="embedding",
+            max_pages=10,
+            
+            # Use OpenAI embeddings
+            embedding_llm_config={
+                'provider': 'openai/text-embedding-3-small',
+                'api_token': os.getenv('OPENAI_API_KEY')
+            },
+            
+            # OpenAI embeddings are high quality, can be stricter
+            embedding_k_exp=4.0,
+            n_query_variations=12
+        )
+        
+        await test_configuration(
+            "OpenAI Embeddings",
+            config_openai,
+            test_url,
+            "event-driven architecture patterns"
+        )
+    
+    # Parameter Guide
+    print("\n" + "="*60)
+    print("PARAMETER TUNING GUIDE")
+    print("="*60)
+    
+    print("\n📊 Key Parameters and Their Effects:")
+    print("\n1. embedding_k_exp (default: 3.0)")
+    print("   - Lower (1-2): More lenient, faster convergence")
+    print("   - Higher (4-5): Stricter, better precision")
+    
+    print("\n2. embedding_coverage_radius (default: 0.2)")
+    print("   - Lower (0.1-0.15): Requires closer matches")
+    print("   - Higher (0.25-0.3): Accepts broader matches")
+    
+    print("\n3. n_query_variations (default: 10)")
+    print("   - Lower (5-7): Faster, less comprehensive")
+    print("   - Higher (15-20): Better coverage, slower")
+    
+    print("\n4. embedding_min_confidence_threshold (default: 0.1)")
+    print("   - Set to 0.15-0.2 for aggressive irrelevance detection")
+    print("   - Set to 0.05 to crawl even barely relevant content")
+    
+    print("\n5. embedding_validation_min_score (default: 0.3)")
+    print("   - Higher (0.5-0.6): Requires strong validation")
+    print("   - Lower (0.2): More permissive stopping")
+    
+    print("\n💡 Tips:")
+    print("- For research: High k_exp, more variations, strict validation")
+    print("- For exploration: Low k_exp, fewer variations, relaxed thresholds")
+    print("- For quality: Focus on overlap_threshold and validation scores")
+    print("- For speed: Reduce variations, increase min_relative_improvement")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/embedding_strategy.py

@@ -0,0 +1,109 @@
+"""
+Embedding Strategy Example for Adaptive Crawling
+
+This example demonstrates how to use the embedding-based strategy
+for semantic understanding and intelligent crawling.
+"""
+
+import asyncio
+import os
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+
+async def main():
+    """Demonstrate embedding strategy for adaptive crawling"""
+    
+    # Configure embedding strategy
+    config = AdaptiveConfig(
+        strategy="embedding",  # Use embedding strategy
+        embedding_model="sentence-transformers/all-MiniLM-L6-v2",  # Default model
+        n_query_variations=10,  # Generate 10 semantic variations
+        max_pages=15,
+        top_k_links=3,
+        min_gain_threshold=0.05,
+        
+        # Embedding-specific parameters
+        embedding_k_exp=3.0,  # Higher = stricter similarity requirements
+        embedding_min_confidence_threshold=0.1,  # Stop if <10% relevant
+        embedding_validation_min_score=0.4  # Validation threshold
+    )
+    
+    # Optional: Use OpenAI embeddings instead
+    if os.getenv('OPENAI_API_KEY'):
+        config.embedding_llm_config = {
+            'provider': 'openai/text-embedding-3-small',
+            'api_token': os.getenv('OPENAI_API_KEY')
+        }
+        print("Using OpenAI embeddings")
+    else:
+        print("Using sentence-transformers (local embeddings)")
+    
+    async with AsyncWebCrawler(verbose=True) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        # Test 1: Relevant query with semantic understanding
+        print("\n" + "="*50)
+        print("TEST 1: Semantic Query Understanding")
+        print("="*50)
+        
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/library/asyncio.html",
+            query="concurrent programming event-driven architecture"
+        )
+        
+        print("\nQuery Expansion:")
+        print(f"Original query expanded to {len(result.expanded_queries)} variations")
+        for i, q in enumerate(result.expanded_queries[:3], 1):
+            print(f"  {i}. {q}")
+        print("  ...")
+        
+        print("\nResults:")
+        adaptive.print_stats(detailed=False)
+        
+        # Test 2: Detecting irrelevant queries
+        print("\n" + "="*50)
+        print("TEST 2: Irrelevant Query Detection")
+        print("="*50)
+        
+        # Reset crawler for new query
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/library/asyncio.html",
+            query="how to bake chocolate chip cookies"
+        )
+        
+        if result.metrics.get('is_irrelevant', False):
+            print("\n✅ Successfully detected irrelevant query!")
+            print(f"Stopped after just {len(result.crawled_urls)} pages")
+            print(f"Reason: {result.metrics.get('stopped_reason', 'unknown')}")
+        else:
+            print("\n❌ Failed to detect irrelevance")
+        
+        print(f"Final confidence: {adaptive.confidence:.1%}")
+        
+        # Test 3: Semantic gap analysis
+        print("\n" + "="*50)
+        print("TEST 3: Semantic Gap Analysis")
+        print("="*50)
+        
+        # Show how embedding strategy identifies gaps
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        result = await adaptive.digest(
+            start_url="https://realpython.com",
+            query="python decorators advanced patterns"
+        )
+        
+        print(f"\nSemantic gaps identified: {len(result.semantic_gaps)}")
+        print(f"Knowledge base embeddings shape: {result.kb_embeddings.shape if result.kb_embeddings is not None else 'None'}")
+        
+        # Show coverage metrics specific to embedding strategy
+        print("\nEmbedding-specific metrics:")
+        print(f"  Average best similarity: {result.metrics.get('avg_best_similarity', 0):.3f}")
+        print(f"  Coverage score: {result.metrics.get('coverage_score', 0):.3f}")
+        print(f"  Validation confidence: {result.metrics.get('validation_confidence', 0):.2%}")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/embedding_vs_statistical.py

@@ -0,0 +1,167 @@
+"""
+Comparison: Embedding vs Statistical Strategy
+
+This example demonstrates the differences between statistical and embedding
+strategies for adaptive crawling, showing when to use each approach.
+"""
+
+import asyncio
+import time
+import os
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+
+async def crawl_with_strategy(url: str, query: str, strategy: str, **kwargs):
+    """Helper function to crawl with a specific strategy"""
+    config = AdaptiveConfig(
+        strategy=strategy,
+        max_pages=20,
+        top_k_links=3,
+        min_gain_threshold=0.05,
+        **kwargs
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        start_time = time.time()
+        result = await adaptive.digest(start_url=url, query=query)
+        elapsed = time.time() - start_time
+        
+        return {
+            'result': result,
+            'crawler': adaptive,
+            'elapsed': elapsed,
+            'pages': len(result.crawled_urls),
+            'confidence': adaptive.confidence
+        }
+
+
+async def main():
+    """Compare embedding and statistical strategies"""
+    
+    # Test scenarios
+    test_cases = [
+        {
+            'name': 'Technical Documentation (Specific Terms)',
+            'url': 'https://docs.python.org/3/library/asyncio.html',
+            'query': 'asyncio.create_task event_loop.run_until_complete'
+        },
+        {
+            'name': 'Conceptual Query (Semantic Understanding)',
+            'url': 'https://docs.python.org/3/library/asyncio.html',
+            'query': 'concurrent programming patterns'
+        },
+        {
+            'name': 'Ambiguous Query',
+            'url': 'https://realpython.com',
+            'query': 'python performance optimization'
+        }
+    ]
+    
+    # Configure embedding strategy
+    embedding_config = {}
+    if os.getenv('OPENAI_API_KEY'):
+        embedding_config['embedding_llm_config'] = {
+            'provider': 'openai/text-embedding-3-small',
+            'api_token': os.getenv('OPENAI_API_KEY')
+        }
+    
+    for test in test_cases:
+        print("\n" + "="*70)
+        print(f"TEST: {test['name']}")
+        print(f"URL: {test['url']}")
+        print(f"Query: '{test['query']}'")
+        print("="*70)
+        
+        # Run statistical strategy
+        print("\n📊 Statistical Strategy:")
+        stat_result = await crawl_with_strategy(
+            test['url'], 
+            test['query'], 
+            'statistical'
+        )
+        
+        print(f"  Pages crawled: {stat_result['pages']}")
+        print(f"  Time taken: {stat_result['elapsed']:.2f}s")
+        print(f"  Confidence: {stat_result['confidence']:.1%}")
+        print(f"  Sufficient: {'Yes' if stat_result['crawler'].is_sufficient else 'No'}")
+        
+        # Show term coverage
+        if hasattr(stat_result['result'], 'term_frequencies'):
+            query_terms = test['query'].lower().split()
+            covered = sum(1 for term in query_terms 
+                         if term in stat_result['result'].term_frequencies)
+            print(f"  Term coverage: {covered}/{len(query_terms)} query terms found")
+        
+        # Run embedding strategy
+        print("\n🧠 Embedding Strategy:")
+        emb_result = await crawl_with_strategy(
+            test['url'], 
+            test['query'], 
+            'embedding',
+            **embedding_config
+        )
+        
+        print(f"  Pages crawled: {emb_result['pages']}")
+        print(f"  Time taken: {emb_result['elapsed']:.2f}s")
+        print(f"  Confidence: {emb_result['confidence']:.1%}")
+        print(f"  Sufficient: {'Yes' if emb_result['crawler'].is_sufficient else 'No'}")
+        
+        # Show semantic understanding
+        if emb_result['result'].expanded_queries:
+            print(f"  Query variations: {len(emb_result['result'].expanded_queries)}")
+            print(f"  Semantic gaps: {len(emb_result['result'].semantic_gaps)}")
+        
+        # Compare results
+        print("\n📈 Comparison:")
+        efficiency_diff = ((stat_result['pages'] - emb_result['pages']) / 
+                          stat_result['pages'] * 100) if stat_result['pages'] > 0 else 0
+        
+        print(f"  Efficiency: ", end="")
+        if efficiency_diff > 0:
+            print(f"Embedding used {efficiency_diff:.0f}% fewer pages")
+        else:
+            print(f"Statistical used {-efficiency_diff:.0f}% fewer pages")
+        
+        print(f"  Speed: ", end="")
+        if stat_result['elapsed'] < emb_result['elapsed']:
+            print(f"Statistical was {emb_result['elapsed']/stat_result['elapsed']:.1f}x faster")
+        else:
+            print(f"Embedding was {stat_result['elapsed']/emb_result['elapsed']:.1f}x faster")
+        
+        print(f"  Confidence difference: {abs(stat_result['confidence'] - emb_result['confidence'])*100:.0f} percentage points")
+        
+        # Recommendation
+        print("\n💡 Recommendation:")
+        if 'specific' in test['name'].lower() or all(len(term) > 5 for term in test['query'].split()):
+            print("  → Statistical strategy is likely better for this use case (specific terms)")
+        elif 'conceptual' in test['name'].lower() or 'semantic' in test['name'].lower():
+            print("  → Embedding strategy is likely better for this use case (semantic understanding)")
+        else:
+            if emb_result['confidence'] > stat_result['confidence'] + 0.1:
+                print("  → Embedding strategy achieved significantly better understanding")
+            elif stat_result['elapsed'] < emb_result['elapsed'] / 2:
+                print("  → Statistical strategy is much faster with similar results")
+            else:
+                print("  → Both strategies performed similarly; choose based on your priorities")
+    
+    # Summary recommendations
+    print("\n" + "="*70)
+    print("STRATEGY SELECTION GUIDE")
+    print("="*70)
+    print("\n✅ Use STATISTICAL strategy when:")
+    print("  - Queries contain specific technical terms")
+    print("  - Speed is critical")
+    print("  - No API access available")
+    print("  - Working with well-structured documentation")
+    
+    print("\n✅ Use EMBEDDING strategy when:")
+    print("  - Queries are conceptual or ambiguous")
+    print("  - Semantic understanding is important")
+    print("  - Need to detect irrelevant content")
+    print("  - Working with diverse content sources")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/adaptive_crawling/export_import_kb.py

@@ -0,0 +1,232 @@
+"""
+Knowledge Base Export and Import
+
+This example demonstrates how to export crawled knowledge bases and
+import them for reuse, sharing, or analysis.
+"""
+
+import asyncio
+import json
+from pathlib import Path
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+
+async def build_knowledge_base():
+    """Build a knowledge base about web technologies"""
+    print("="*60)
+    print("PHASE 1: Building Knowledge Base")
+    print("="*60)
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler)
+        
+        # Crawl information about HTTP
+        print("\n1. Gathering HTTP protocol information...")
+        await adaptive.digest(
+            start_url="https://httpbin.org",
+            query="http methods headers status codes"
+        )
+        print(f"   - Pages crawled: {len(adaptive.state.crawled_urls)}")
+        print(f"   - Confidence: {adaptive.confidence:.2%}")
+        
+        # Add more information about APIs
+        print("\n2. Adding API documentation knowledge...")
+        await adaptive.digest(
+            start_url="https://httpbin.org/anything",
+            query="rest api json response request"
+        )
+        print(f"   - Total pages: {len(adaptive.state.crawled_urls)}")
+        print(f"   - Confidence: {adaptive.confidence:.2%}")
+        
+        # Export the knowledge base
+        export_path = "web_tech_knowledge.jsonl"
+        print(f"\n3. Exporting knowledge base to {export_path}")
+        adaptive.export_knowledge_base(export_path)
+        
+        # Show export statistics
+        export_size = Path(export_path).stat().st_size / 1024
+        with open(export_path, 'r') as f:
+            line_count = sum(1 for _ in f)
+        
+        print(f"   - Exported {line_count} documents")
+        print(f"   - File size: {export_size:.1f} KB")
+        
+        return export_path
+
+
+async def analyze_knowledge_base(kb_path):
+    """Analyze the exported knowledge base"""
+    print("\n" + "="*60)
+    print("PHASE 2: Analyzing Exported Knowledge Base")
+    print("="*60)
+    
+    # Read and analyze JSONL
+    documents = []
+    with open(kb_path, 'r') as f:
+        for line in f:
+            documents.append(json.loads(line))
+    
+    print(f"\nKnowledge base contains {len(documents)} documents:")
+    
+    # Analyze document properties
+    total_content_length = 0
+    urls_by_domain = {}
+    
+    for doc in documents:
+        # Content analysis
+        content_length = len(doc.get('content', ''))
+        total_content_length += content_length
+        
+        # URL analysis
+        url = doc.get('url', '')
+        domain = url.split('/')[2] if url.startswith('http') else 'unknown'
+        urls_by_domain[domain] = urls_by_domain.get(domain, 0) + 1
+        
+        # Show sample document
+        if documents.index(doc) == 0:
+            print(f"\nSample document structure:")
+            print(f"  - URL: {url}")
+            print(f"  - Content length: {content_length} chars")
+            print(f"  - Has metadata: {'metadata' in doc}")
+            print(f"  - Has links: {len(doc.get('links', []))} links")
+            print(f"  - Query: {doc.get('query', 'N/A')}")
+    
+    print(f"\nContent statistics:")
+    print(f"  - Total content: {total_content_length:,} characters")
+    print(f"  - Average per document: {total_content_length/len(documents):,.0f} chars")
+    
+    print(f"\nDomain distribution:")
+    for domain, count in urls_by_domain.items():
+        print(f"  - {domain}: {count} pages")
+
+
+async def import_and_continue():
+    """Import a knowledge base and continue crawling"""
+    print("\n" + "="*60)
+    print("PHASE 3: Importing and Extending Knowledge Base")
+    print("="*60)
+    
+    kb_path = "web_tech_knowledge.jsonl"
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        # Create new adaptive crawler
+        adaptive = AdaptiveCrawler(crawler)
+        
+        # Import existing knowledge base
+        print(f"\n1. Importing knowledge base from {kb_path}")
+        adaptive.import_knowledge_base(kb_path)
+        
+        print(f"   - Imported {len(adaptive.state.knowledge_base)} documents")
+        print(f"   - Existing URLs: {len(adaptive.state.crawled_urls)}")
+        
+        # Check current state
+        print("\n2. Checking imported knowledge state:")
+        adaptive.print_stats(detailed=False)
+        
+        # Continue crawling with new query
+        print("\n3. Extending knowledge with new query...")
+        await adaptive.digest(
+            start_url="https://httpbin.org/status/200",
+            query="error handling retry timeout"
+        )
+        
+        print("\n4. Final knowledge base state:")
+        adaptive.print_stats(detailed=False)
+        
+        # Export extended knowledge base
+        extended_path = "web_tech_knowledge_extended.jsonl"
+        adaptive.export_knowledge_base(extended_path)
+        print(f"\n5. Extended knowledge base exported to {extended_path}")
+
+
+async def share_knowledge_bases():
+    """Demonstrate sharing knowledge bases between projects"""
+    print("\n" + "="*60)
+    print("PHASE 4: Sharing Knowledge Between Projects")
+    print("="*60)
+    
+    # Simulate two different projects
+    project_a_kb = "project_a_knowledge.jsonl"
+    project_b_kb = "project_b_knowledge.jsonl"
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        # Project A: Security documentation
+        print("\n1. Project A: Building security knowledge...")
+        crawler_a = AdaptiveCrawler(crawler)
+        await crawler_a.digest(
+            start_url="https://httpbin.org/basic-auth/user/pass",
+            query="authentication security headers"
+        )
+        crawler_a.export_knowledge_base(project_a_kb)
+        print(f"   - Exported {len(crawler_a.state.knowledge_base)} documents")
+        
+        # Project B: API testing
+        print("\n2. Project B: Building testing knowledge...")
+        crawler_b = AdaptiveCrawler(crawler)
+        await crawler_b.digest(
+            start_url="https://httpbin.org/anything",
+            query="testing endpoints mocking"
+        )
+        crawler_b.export_knowledge_base(project_b_kb)
+        print(f"   - Exported {len(crawler_b.state.knowledge_base)} documents")
+        
+        # Merge knowledge bases
+        print("\n3. Merging knowledge bases...")
+        merged_crawler = AdaptiveCrawler(crawler)
+        
+        # Import both knowledge bases
+        merged_crawler.import_knowledge_base(project_a_kb)
+        initial_size = len(merged_crawler.state.knowledge_base)
+        
+        merged_crawler.import_knowledge_base(project_b_kb)
+        final_size = len(merged_crawler.state.knowledge_base)
+        
+        print(f"   - Project A documents: {initial_size}")
+        print(f"   - Additional from Project B: {final_size - initial_size}")
+        print(f"   - Total merged documents: {final_size}")
+        
+        # Export merged knowledge
+        merged_kb = "merged_knowledge.jsonl"
+        merged_crawler.export_knowledge_base(merged_kb)
+        print(f"\n4. Merged knowledge base exported to {merged_kb}")
+        
+        # Show combined coverage
+        print("\n5. Combined knowledge coverage:")
+        merged_crawler.print_stats(detailed=False)
+
+
+async def main():
+    """Run all examples"""
+    try:
+        # Build initial knowledge base
+        kb_path = await build_knowledge_base()
+        
+        # Analyze the export
+        await analyze_knowledge_base(kb_path)
+        
+        # Import and extend
+        await import_and_continue()
+        
+        # Demonstrate sharing
+        await share_knowledge_bases()
+        
+        print("\n" + "="*60)
+        print("All examples completed successfully!")
+        print("="*60)
+        
+    finally:
+        # Clean up generated files
+        print("\nCleaning up generated files...")
+        for file in [
+            "web_tech_knowledge.jsonl",
+            "web_tech_knowledge_extended.jsonl", 
+            "project_a_knowledge.jsonl",
+            "project_b_knowledge.jsonl",
+            "merged_knowledge.jsonl"
+        ]:
+            Path(file).unlink(missing_ok=True)
+        print("Cleanup complete.")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/c4a_script/api_usage_examples.py

@@ -3,8 +3,8 @@ C4A-Script API Usage Examples
 Shows how to use the new Result-based API in various scenarios
 """
 
-from c4a_compile import compile, validate, compile_file
-from c4a_result import CompilationResult, ValidationResult
+from crawl4ai.script.c4a_compile import compile, validate, compile_file
+from crawl4ai.script.c4a_result import CompilationResult, ValidationResult
 import json
 
 

docs/examples/c4a_script/c4a_script_hello_world.py

@@ -3,7 +3,7 @@ C4A-Script Hello World
 A concise example showing how to use the C4A-Script compiler
 """
 
-from c4a_compile import compile
+from crawl4ai.script.c4a_compile import compile
 
 # Define your C4A-Script
 script = """

docs/examples/c4a_script/c4a_script_hello_world_error.py

@@ -3,7 +3,7 @@ C4A-Script Hello World - Error Example
 Shows how error handling works
 """
 
-from c4a_compile import compile
+from crawl4ai.script.c4a_compile import compile
 
 # Define a script with an error (missing THEN)
 script = """

docs/examples/demo_multi_config_clean.py

@@ -0,0 +1,303 @@
+"""
+🎯 Multi-Config URL Matching Demo
+=================================
+Learn how to use different crawler configurations for different URL patterns
+in a single crawl batch with Crawl4AI's multi-config feature.
+
+Part 1: Understanding URL Matching (Pattern Testing)
+Part 2: Practical Example with Real Crawling
+"""
+
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler, 
+    CrawlerRunConfig,
+    MatchMode
+)
+from crawl4ai.processors.pdf import PDFContentScrapingStrategy
+from crawl4ai.extraction_strategy import JsonCssExtractionStrategy
+from crawl4ai.content_filter_strategy import PruningContentFilter
+from crawl4ai.markdown_generation_strategy import DefaultMarkdownGenerator
+
+
+def print_section(title):
+    """Print a formatted section header"""
+    print(f"\n{'=' * 60}")
+    print(f"{title}")
+    print(f"{'=' * 60}\n")
+
+
+def test_url_matching(config, test_urls, config_name):
+    """Test URL matching for a config and show results"""
+    print(f"Config: {config_name}")
+    print(f"Matcher: {config.url_matcher}")
+    if hasattr(config, 'match_mode'):
+        print(f"Mode: {config.match_mode.value}")
+    print("-" * 40)
+    
+    for url in test_urls:
+        matches = config.is_match(url)
+        symbol = "✓" if matches else "✗"
+        print(f"{symbol} {url}")
+    print()
+
+
+# ==============================================================================
+# PART 1: Understanding URL Matching
+# ==============================================================================
+
+def demo_part1_pattern_matching():
+    """Part 1: Learn how URL matching works without crawling"""
+    
+    print_section("PART 1: Understanding URL Matching")
+    print("Let's explore different ways to match URLs with configs.\n")
+    
+    # Test URLs we'll use throughout
+    test_urls = [
+        "https://example.com/report.pdf",
+        "https://example.com/data.json",
+        "https://example.com/blog/post-1",
+        "https://example.com/article/news",
+        "https://api.example.com/v1/users",
+        "https://example.com/about"
+    ]
+    
+    # 1.1 Simple String Pattern
+    print("1.1 Simple String Pattern Matching")
+    print("-" * 40)
+    
+    pdf_config = CrawlerRunConfig(
+        url_matcher="*.pdf"
+    )
+    
+    test_url_matching(pdf_config, test_urls, "PDF Config")
+    
+    
+    # 1.2 Multiple String Patterns
+    print("1.2 Multiple String Patterns (OR logic)")
+    print("-" * 40)
+    
+    blog_config = CrawlerRunConfig(
+        url_matcher=["*/blog/*", "*/article/*", "*/news/*"],
+        match_mode=MatchMode.OR  # This is default, shown for clarity
+    )
+    
+    test_url_matching(blog_config, test_urls, "Blog/Article Config")
+    
+    
+    # 1.3 Single Function Matcher
+    print("1.3 Function-based Matching")
+    print("-" * 40)
+    
+    api_config = CrawlerRunConfig(
+        url_matcher=lambda url: 'api' in url or url.endswith('.json')
+    )
+    
+    test_url_matching(api_config, test_urls, "API Config")
+    
+    
+    # 1.4 List of Functions
+    print("1.4 Multiple Functions with AND Logic")
+    print("-" * 40)
+    
+    # Must be HTTPS AND contain 'api' AND have version number
+    secure_api_config = CrawlerRunConfig(
+        url_matcher=[
+            lambda url: url.startswith('https://'),
+            lambda url: 'api' in url,
+            lambda url: '/v' in url  # Version indicator
+        ],
+        match_mode=MatchMode.AND
+    )
+    
+    test_url_matching(secure_api_config, test_urls, "Secure API Config")
+    
+    
+    # 1.5 Mixed: String and Function Together
+    print("1.5 Mixed Patterns: String + Function")
+    print("-" * 40)
+    
+    # Match JSON files OR any API endpoint
+    json_or_api_config = CrawlerRunConfig(
+        url_matcher=[
+            "*.json",  # String pattern
+            lambda url: 'api' in url  # Function
+        ],
+        match_mode=MatchMode.OR
+    )
+    
+    test_url_matching(json_or_api_config, test_urls, "JSON or API Config")
+    
+    
+    # 1.6 Complex: Multiple Strings + Multiple Functions
+    print("1.6 Complex Matcher: Mixed Types with AND Logic")
+    print("-" * 40)
+    
+    # Must be: HTTPS AND (.com domain) AND (blog OR article) AND NOT a PDF
+    complex_config = CrawlerRunConfig(
+        url_matcher=[
+            lambda url: url.startswith('https://'),  # Function: HTTPS check
+            "*.com/*",  # String: .com domain
+            lambda url: any(pattern in url for pattern in ['/blog/', '/article/']),  # Function: Blog OR article
+            lambda url: not url.endswith('.pdf')  # Function: Not PDF
+        ],
+        match_mode=MatchMode.AND
+    )
+    
+    test_url_matching(complex_config, test_urls, "Complex Mixed Config")
+    
+    print("\n✅ Key Takeaway: First matching config wins when passed to arun_many()!")
+
+
+# ==============================================================================
+# PART 2: Practical Multi-URL Crawling
+# ==============================================================================
+
+async def demo_part2_practical_crawling():
+    """Part 2: Real-world example with different content types"""
+    
+    print_section("PART 2: Practical Multi-URL Crawling")
+    print("Now let's see multi-config in action with real URLs.\n")
+    
+    # Create specialized configs for different content types
+    configs = [
+        # Config 1: PDF documents - only match files ending with .pdf
+        CrawlerRunConfig(
+            url_matcher="*.pdf",
+            scraping_strategy=PDFContentScrapingStrategy()
+        ),
+        
+        # Config 2: Blog/article pages with content filtering
+        CrawlerRunConfig(
+            url_matcher=["*/blog/*", "*/article/*", "*python.org*"],
+            markdown_generator=DefaultMarkdownGenerator(
+                content_filter=PruningContentFilter(threshold=0.48)
+            )
+        ),
+        
+        # Config 3: Dynamic pages requiring JavaScript
+        CrawlerRunConfig(
+            url_matcher=lambda url: 'github.com' in url,
+            js_code="window.scrollTo(0, 500);"  # Scroll to load content
+        ),
+        
+        # Config 4: Mixed matcher - API endpoints (string OR function)
+        CrawlerRunConfig(
+            url_matcher=[
+                "*.json",  # String pattern for JSON files
+                lambda url: 'api' in url or 'httpbin.org' in url  # Function for API endpoints
+            ],
+            match_mode=MatchMode.OR,
+        ),
+        
+        # Config 5: Complex matcher - Secure documentation sites
+        CrawlerRunConfig(
+            url_matcher=[
+                lambda url: url.startswith('https://'),  # Must be HTTPS
+                "*.org/*",  # String: .org domain
+                lambda url: any(doc in url for doc in ['docs', 'documentation', 'reference']),  # Has docs
+                lambda url: not url.endswith(('.pdf', '.json'))  # Not PDF or JSON
+            ],
+            match_mode=MatchMode.AND,
+            # wait_for="css:.content, css:article"  # Wait for content to load
+        ),
+        
+        # Default config for everything else
+        # CrawlerRunConfig()  # No url_matcher means it matches everything (use it as fallback)
+    ]
+    
+    # URLs to crawl - each will use a different config
+    urls = [
+        "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf",  # → PDF config
+        "https://blog.python.org/",  # → Blog config with content filter
+        "https://github.com/microsoft/playwright",  # → JS config
+        "https://httpbin.org/json",  # → Mixed matcher config (API)
+        "https://docs.python.org/3/reference/",  # → Complex matcher config
+        "https://www.w3schools.com/",  # → Default config, if you uncomment the default config line above, if not you will see `Error: No matching configuration`
+    ]
+    
+    print("URLs to crawl:")
+    for i, url in enumerate(urls, 1):
+        print(f"{i}. {url}")
+    
+    print("\nCrawling with appropriate config for each URL...\n")
+    
+    async with AsyncWebCrawler() as crawler:
+        results = await crawler.arun_many(
+            urls=urls,
+            config=configs
+        )
+        
+        # Display results
+        print("Results:")
+        print("-" * 60)
+        
+        for result in results:
+            if result.success:
+                # Determine which config was used
+                config_type = "Default"
+                if result.url.endswith('.pdf'):
+                    config_type = "PDF Strategy"
+                elif any(pattern in result.url for pattern in ['blog', 'python.org']) and 'docs' not in result.url:
+                    config_type = "Blog + Content Filter"
+                elif 'github.com' in result.url:
+                    config_type = "JavaScript Enabled"
+                elif 'httpbin.org' in result.url or result.url.endswith('.json'):
+                    config_type = "Mixed Matcher (API)"
+                elif 'docs.python.org' in result.url:
+                    config_type = "Complex Matcher (Secure Docs)"
+                
+                print(f"\n✓ {result.url}")
+                print(f"  Config used: {config_type}")
+                print(f"  Content size: {len(result.markdown)} chars")
+                
+                # Show if we have fit_markdown (from content filter)
+                if hasattr(result.markdown, 'fit_markdown') and result.markdown.fit_markdown:
+                    print(f"  Fit markdown size: {len(result.markdown.fit_markdown)} chars")
+                    reduction = (1 - len(result.markdown.fit_markdown) / len(result.markdown)) * 100
+                    print(f"  Content reduced by: {reduction:.1f}%")
+                
+                # Show extracted data if using extraction strategy
+                if hasattr(result, 'extracted_content') and result.extracted_content:
+                    print(f"  Extracted data: {str(result.extracted_content)[:100]}...")
+            else:
+                print(f"\n✗ {result.url}")
+                print(f"  Error: {result.error_message}")
+    
+    print("\n" + "=" * 60)
+    print("✅ Multi-config crawling complete!")
+    print("\nBenefits demonstrated:")
+    print("- PDFs handled with specialized scraper")
+    print("- Blog content filtered for relevance") 
+    print("- JavaScript executed only where needed")
+    print("- Mixed matchers (string + function) for flexible matching")
+    print("- Complex matchers for precise URL targeting")
+    print("- Each URL got optimal configuration automatically!")
+
+
+async def main():
+    """Run both parts of the demo"""
+    
+    print("""
+🎯 Multi-Config URL Matching Demo
+=================================
+Learn how Crawl4AI can use different configurations
+for different URLs in a single batch.
+    """)
+    
+    # Part 1: Pattern matching
+    demo_part1_pattern_matching()
+    
+    print("\nPress Enter to continue to Part 2...")
+    try:
+        input()
+    except EOFError:
+        # Running in non-interactive mode, skip input
+        pass
+    
+    # Part 2: Practical crawling
+    await demo_part2_practical_crawling()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/hello_world_undetected.py

@@ -0,0 +1,57 @@
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler,
+    BrowserConfig,
+    CrawlerRunConfig,
+    DefaultMarkdownGenerator,
+    PruningContentFilter,
+    CrawlResult,
+    UndetectedAdapter
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+
+async def main():
+    # Create browser config
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True,
+    )
+    
+    # Create the undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    
+    # Create the crawler strategy with the undetected adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    # Create the crawler with our custom strategy
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        # Configure the crawl
+        crawler_config = CrawlerRunConfig(
+            markdown_generator=DefaultMarkdownGenerator(
+                content_filter=PruningContentFilter()
+            ),
+            capture_console_messages=True,  # Enable console capture to test adapter
+        )
+        
+        # Test on a site that typically detects bots
+        print("Testing undetected adapter...")
+        result: CrawlResult = await crawler.arun(
+            url="https://www.helloworld.org", 
+            config=crawler_config
+        )
+        
+        print(f"Status: {result.status_code}")
+        print(f"Success: {result.success}")
+        print(f"Console messages captured: {len(result.console_messages or [])}")
+        print(f"Markdown content (first 500 chars):\n{result.markdown.raw_markdown[:500]}")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/link_head_extraction_example.py

@@ -18,7 +18,7 @@ Usage:
 
 import asyncio
 from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
-from crawl4ai.async_configs import LinkPreviewConfig
+from crawl4ai import LinkPreviewConfig
 
 
 async def basic_link_head_extraction():

docs/examples/llm_extraction_openai_pricing.py

@@ -1,43 +1,55 @@
-from crawl4ai import LLMConfig
-from crawl4ai import AsyncWebCrawler, LLMExtractionStrategy
 import asyncio
-import os
-import json
 from pydantic import BaseModel, Field
-
-url = "https://openai.com/api/pricing/"
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, LLMConfig, BrowserConfig, CacheMode
+from crawl4ai.extraction_strategy import LLMExtractionStrategy
+from typing import Dict
+import os
 
 
 class OpenAIModelFee(BaseModel):
     model_name: str = Field(..., description="Name of the OpenAI model.")
     input_fee: str = Field(..., description="Fee for input token for the OpenAI model.")
-    output_fee: str = Field(
-        ..., description="Fee for output token for the OpenAI model."
+    output_fee: str = Field(..., description="Fee for output token for the OpenAI model.")
+
+
+async def extract_structured_data_using_llm(provider: str, api_token: str = None, extra_headers: Dict[str, str] = None):
+    print(f"\n--- Extracting Structured Data with {provider} ---")
+
+    if api_token is None and provider != "ollama":
+        print(f"API token is required for {provider}. Skipping this example.")
+        return
+
+    browser_config = BrowserConfig(headless=True)
+
+    extra_args = {"temperature": 0, "top_p": 0.9, "max_tokens": 2000}
+    if extra_headers:
+        extra_args["extra_headers"] = extra_headers
+
+    crawler_config = CrawlerRunConfig(
+        cache_mode=CacheMode.BYPASS,
+        word_count_threshold=1,
+        page_timeout=80000,
+        extraction_strategy=LLMExtractionStrategy(
+            llm_config=LLMConfig(provider=provider, api_token=api_token),
+            schema=OpenAIModelFee.model_json_schema(),
+            extraction_type="schema",
+            instruction="""From the crawled content, extract all mentioned model names along with their fees for input and output tokens. 
+            Do not miss any models in the entire content.""",
+            extra_args=extra_args,
+        ),
     )
 
-async def main():
-    # Use AsyncWebCrawler
-    async with AsyncWebCrawler() as crawler:
+    async with AsyncWebCrawler(config=browser_config) as crawler:
         result = await crawler.arun(
-            url=url,
-            word_count_threshold=1,
-            extraction_strategy=LLMExtractionStrategy(
-                # provider= "openai/gpt-4o", api_token = os.getenv('OPENAI_API_KEY'),
-                llm_config=LLMConfig(provider="groq/llama-3.1-70b-versatile", api_token=os.getenv("GROQ_API_KEY")),
-                schema=OpenAIModelFee.model_json_schema(),
-                extraction_type="schema",
-                instruction="From the crawled content, extract all mentioned model names along with their "
-                "fees for input and output tokens. Make sure not to miss anything in the entire content. "
-                "One extracted model JSON format should look like this: "
-                '{ "model_name": "GPT-4", "input_fee": "US$10.00 / 1M tokens", "output_fee": "US$30.00 / 1M tokens" }',
-            ),
+            url="https://openai.com/api/pricing/", 
+            config=crawler_config
         )
-        print("Success:", result.success)
-        model_fees = json.loads(result.extracted_content)
-        print(len(model_fees))
-
-        with open(".data/data.json", "w", encoding="utf-8") as f:
-            f.write(result.extracted_content)
+        print(result.extracted_content)
 
 
-asyncio.run(main())
+if __name__ == "__main__":
+    asyncio.run(
+        extract_structured_data_using_llm(
+            provider="openai/gpt-4o", api_token=os.getenv("OPENAI_API_KEY")
+        )
+    )

docs/examples/scraping_strategies_performance.py

@@ -1,5 +1,6 @@
 import time, re
-from crawl4ai.content_scraping_strategy import WebScrapingStrategy,  LXMLWebScrapingStrategy
+from crawl4ai.content_scraping_strategy import LXMLWebScrapingStrategy
+# WebScrapingStrategy is now an alias for LXMLWebScrapingStrategy
 import time
 import functools
 from collections import defaultdict
@@ -57,7 +58,7 @@ methods_to_profile = [
 
 
 # Apply decorators to both strategies
-for strategy, name in [(WebScrapingStrategy, "Original"), (LXMLWebScrapingStrategy, "LXML")]:
+for strategy, name in [(LXMLWebScrapingStrategy, "LXML")]:
     for method in methods_to_profile:
         apply_decorators(strategy, method, name)
 
@@ -85,7 +86,7 @@ def generate_large_html(n_elements=1000):
 
 def test_scraping():
     # Initialize both scrapers
-    original_scraper = WebScrapingStrategy()
+    original_scraper = LXMLWebScrapingStrategy()
     selected_scraper = LXMLWebScrapingStrategy()
     
     # Generate test HTML

docs/examples/simple_anti_bot_examples.py

@@ -0,0 +1,59 @@
+import asyncio
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig, UndetectedAdapter
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+# Example 1: Stealth Mode
+async def stealth_mode_example():
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun("https://example.com")
+        return result.html[:500]
+
+# Example 2: Undetected Browser
+async def undetected_browser_example():
+    browser_config = BrowserConfig(
+        headless=False
+    )
+    
+    adapter = UndetectedAdapter()
+    strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=strategy,
+        config=browser_config
+    ) as crawler:
+        result = await crawler.arun("https://example.com")
+        return result.html[:500]
+
+# Example 3: Both Combined
+async def combined_example():
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False
+    )
+    
+    adapter = UndetectedAdapter()
+    strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=strategy,
+        config=browser_config
+    ) as crawler:
+        result = await crawler.arun("https://example.com")
+        return result.html[:500]
+
+# Run examples
+if __name__ == "__main__":
+    asyncio.run(stealth_mode_example())
+    asyncio.run(undetected_browser_example())
+    asyncio.run(combined_example())

docs/examples/stealth_mode_example.py

@@ -0,0 +1,522 @@
+"""
+Stealth Mode Example with Crawl4AI
+
+This example demonstrates how to use the stealth mode feature to bypass basic bot detection.
+The stealth mode uses playwright-stealth to modify browser fingerprints and behaviors
+that are commonly used to detect automated browsers.
+
+Key features demonstrated:
+1. Comparing crawling with and without stealth mode
+2. Testing against bot detection sites
+3. Accessing sites that block automated browsers
+4. Best practices for stealth crawling
+"""
+
+import asyncio
+import json
+from typing import Dict, Any
+from colorama import Fore, Style, init
+
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
+from crawl4ai.async_logger import AsyncLogger
+
+# Initialize colorama for colored output
+init()
+
+# Create a logger for better output
+logger = AsyncLogger(verbose=True)
+
+
+async def test_bot_detection(use_stealth: bool = False) -> Dict[str, Any]:
+    """Test against a bot detection service"""
+    
+    logger.info(
+        f"Testing bot detection with stealth={'ON' if use_stealth else 'OFF'}",
+        tag="STEALTH"
+    )
+    
+    # Configure browser with or without stealth
+    browser_config = BrowserConfig(
+        headless=False,  # Use False to see the browser in action
+        enable_stealth=use_stealth,
+        viewport_width=1280,
+        viewport_height=800
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        # JavaScript to extract bot detection results
+        detection_script = """
+        // Comprehensive bot detection checks
+        (() => {
+        const detectionResults = {
+            // Basic WebDriver detection
+            webdriver: navigator.webdriver,
+            
+            // Chrome specific
+            chrome: !!window.chrome,
+            chromeRuntime: !!window.chrome?.runtime,
+            
+            // Automation indicators
+            automationControlled: navigator.webdriver,
+            
+            // Permissions API
+            permissionsPresent: !!navigator.permissions?.query,
+            
+            // Plugins
+            pluginsLength: navigator.plugins.length,
+            pluginsArray: Array.from(navigator.plugins).map(p => p.name),
+            
+            // Languages
+            languages: navigator.languages,
+            language: navigator.language,
+            
+            // User agent
+            userAgent: navigator.userAgent,
+            
+            // Screen and window properties
+            screen: {
+                width: screen.width,
+                height: screen.height,
+                availWidth: screen.availWidth,
+                availHeight: screen.availHeight,
+                colorDepth: screen.colorDepth,
+                pixelDepth: screen.pixelDepth
+            },
+            
+            // WebGL vendor
+            webglVendor: (() => {
+                try {
+                    const canvas = document.createElement('canvas');
+                    const gl = canvas.getContext('webgl') || canvas.getContext('experimental-webgl');
+                    const ext = gl.getExtension('WEBGL_debug_renderer_info');
+                    return gl.getParameter(ext.UNMASKED_VENDOR_WEBGL);
+                } catch (e) {
+                    return 'Error';
+                }
+            })(),
+            
+            // Platform
+            platform: navigator.platform,
+            
+            // Hardware concurrency
+            hardwareConcurrency: navigator.hardwareConcurrency,
+            
+            // Device memory
+            deviceMemory: navigator.deviceMemory,
+            
+            // Connection
+            connection: navigator.connection?.effectiveType
+        };
+        
+        // Log results for console capture
+        console.log('DETECTION_RESULTS:', JSON.stringify(detectionResults, null, 2));
+        
+        // Return results
+        return detectionResults;
+        })();
+        """
+        
+        # Crawl bot detection test page
+        config = CrawlerRunConfig(
+            js_code=detection_script,
+            capture_console_messages=True,
+            wait_until="networkidle",
+            delay_before_return_html=2.0  # Give time for all checks to complete
+        )
+        
+        result = await crawler.arun(
+            url="https://bot.sannysoft.com",
+            config=config
+        )
+        
+        if result.success:
+            # Extract detection results from console
+            detection_data = None
+            for msg in result.console_messages or []:
+                if "DETECTION_RESULTS:" in msg.get("text", ""):
+                    try:
+                        json_str = msg["text"].replace("DETECTION_RESULTS:", "").strip()
+                        detection_data = json.loads(json_str)
+                    except:
+                        pass
+            
+            # Also try to get from JavaScript execution result
+            if not detection_data and result.js_execution_result:
+                detection_data = result.js_execution_result
+            
+            return {
+                "success": True,
+                "url": result.url,
+                "detection_data": detection_data,
+                "page_title": result.metadata.get("title", ""),
+                "stealth_enabled": use_stealth
+            }
+        else:
+            return {
+                "success": False,
+                "error": result.error_message,
+                "stealth_enabled": use_stealth
+            }
+
+
+async def test_cloudflare_site(use_stealth: bool = False) -> Dict[str, Any]:
+    """Test accessing a Cloudflare-protected site"""
+    
+    logger.info(
+        f"Testing Cloudflare site with stealth={'ON' if use_stealth else 'OFF'}",
+        tag="STEALTH"
+    )
+    
+    browser_config = BrowserConfig(
+        headless=True,  # Cloudflare detection works better in headless mode with stealth
+        enable_stealth=use_stealth,
+        viewport_width=1920,
+        viewport_height=1080
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        config = CrawlerRunConfig(
+            wait_until="networkidle",
+            page_timeout=30000,  # 30 seconds
+            delay_before_return_html=3.0
+        )
+        
+        # Test on a site that often shows Cloudflare challenges
+        result = await crawler.arun(
+            url="https://nowsecure.nl",
+            config=config
+        )
+        
+        # Check if we hit Cloudflare challenge
+        cloudflare_detected = False
+        if result.html:
+            cloudflare_indicators = [
+                "Checking your browser",
+                "Just a moment",
+                "cf-browser-verification",
+                "cf-challenge",
+                "ray ID"
+            ]
+            cloudflare_detected = any(indicator in result.html for indicator in cloudflare_indicators)
+        
+        return {
+            "success": result.success,
+            "url": result.url,
+            "cloudflare_challenge": cloudflare_detected,
+            "status_code": result.status_code,
+            "page_title": result.metadata.get("title", "") if result.metadata else "",
+            "stealth_enabled": use_stealth,
+            "html_snippet": result.html[:500] if result.html else ""
+        }
+
+
+async def test_anti_bot_site(use_stealth: bool = False) -> Dict[str, Any]:
+    """Test against sites with anti-bot measures"""
+    
+    logger.info(
+        f"Testing anti-bot site with stealth={'ON' if use_stealth else 'OFF'}",
+        tag="STEALTH"
+    )
+    
+    browser_config = BrowserConfig(
+        headless=False,
+        enable_stealth=use_stealth,
+        # Additional browser arguments that help with stealth
+        extra_args=[
+            "--disable-blink-features=AutomationControlled",
+            "--disable-features=site-per-process"
+        ] if not use_stealth else []  # These are automatically applied with stealth
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        # Some sites check for specific behaviors
+        behavior_script = """
+        (async () => {
+            // Simulate human-like behavior
+            const sleep = ms => new Promise(resolve => setTimeout(resolve, ms));
+            
+            // Random mouse movement
+            const moveX = Math.random() * 100;
+            const moveY = Math.random() * 100;
+            
+            // Simulate reading time
+            await sleep(1000 + Math.random() * 2000);
+            
+            // Scroll slightly
+            window.scrollBy(0, 100 + Math.random() * 200);
+            
+            console.log('Human behavior simulation complete');
+            return true;
+        })()
+        """
+        
+        config = CrawlerRunConfig(
+            js_code=behavior_script,
+            wait_until="networkidle",
+            delay_before_return_html=5.0,  # Longer delay to appear more human
+            capture_console_messages=True
+        )
+        
+        # Test on a site that implements anti-bot measures
+        result = await crawler.arun(
+            url="https://www.g2.com/",
+            config=config
+        )
+        
+        # Check for common anti-bot blocks
+        blocked_indicators = [
+            "Access Denied",
+            "403 Forbidden", 
+            "Security Check",
+            "Verify you are human",
+            "captcha",
+            "challenge"
+        ]
+        
+        blocked = False
+        if result.html:
+            blocked = any(indicator.lower() in result.html.lower() for indicator in blocked_indicators)
+        
+        return {
+            "success": result.success and not blocked,
+            "url": result.url,
+            "blocked": blocked,
+            "status_code": result.status_code,
+            "page_title": result.metadata.get("title", "") if result.metadata else "",
+            "stealth_enabled": use_stealth
+        }
+
+
+async def compare_results():
+    """Run all tests with and without stealth mode and compare results"""
+    
+    print(f"\n{Fore.CYAN}{'='*60}{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}Crawl4AI Stealth Mode Comparison{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}{'='*60}{Style.RESET_ALL}\n")
+    
+    # Test 1: Bot Detection
+    print(f"{Fore.YELLOW}1. Bot Detection Test (bot.sannysoft.com){Style.RESET_ALL}")
+    print("-" * 40)
+    
+    # Without stealth
+    regular_detection = await test_bot_detection(use_stealth=False)
+    if regular_detection["success"] and regular_detection["detection_data"]:
+        print(f"{Fore.RED}Without Stealth:{Style.RESET_ALL}")
+        data = regular_detection["detection_data"]
+        print(f"  • WebDriver detected: {data.get('webdriver', 'Unknown')}")
+        print(f"  • Chrome: {data.get('chrome', 'Unknown')}")
+        print(f"  • Languages: {data.get('languages', 'Unknown')}")
+        print(f"  • Plugins: {data.get('pluginsLength', 'Unknown')}")
+        print(f"  • User Agent: {data.get('userAgent', 'Unknown')[:60]}...")
+    
+    # With stealth
+    stealth_detection = await test_bot_detection(use_stealth=True)
+    if stealth_detection["success"] and stealth_detection["detection_data"]:
+        print(f"\n{Fore.GREEN}With Stealth:{Style.RESET_ALL}")
+        data = stealth_detection["detection_data"]
+        print(f"  • WebDriver detected: {data.get('webdriver', 'Unknown')}")
+        print(f"  • Chrome: {data.get('chrome', 'Unknown')}")
+        print(f"  • Languages: {data.get('languages', 'Unknown')}")
+        print(f"  • Plugins: {data.get('pluginsLength', 'Unknown')}")
+        print(f"  • User Agent: {data.get('userAgent', 'Unknown')[:60]}...")
+    
+    # Test 2: Cloudflare Site
+    print(f"\n\n{Fore.YELLOW}2. Cloudflare Protected Site Test{Style.RESET_ALL}")
+    print("-" * 40)
+    
+    # Without stealth
+    regular_cf = await test_cloudflare_site(use_stealth=False)
+    print(f"{Fore.RED}Without Stealth:{Style.RESET_ALL}")
+    print(f"  • Success: {regular_cf['success']}")
+    print(f"  • Cloudflare Challenge: {regular_cf['cloudflare_challenge']}")
+    print(f"  • Status Code: {regular_cf['status_code']}")
+    print(f"  • Page Title: {regular_cf['page_title']}")
+    
+    # With stealth
+    stealth_cf = await test_cloudflare_site(use_stealth=True)
+    print(f"\n{Fore.GREEN}With Stealth:{Style.RESET_ALL}")
+    print(f"  • Success: {stealth_cf['success']}")
+    print(f"  • Cloudflare Challenge: {stealth_cf['cloudflare_challenge']}")
+    print(f"  • Status Code: {stealth_cf['status_code']}")
+    print(f"  • Page Title: {stealth_cf['page_title']}")
+    
+    # Test 3: Anti-bot Site
+    print(f"\n\n{Fore.YELLOW}3. Anti-Bot Site Test{Style.RESET_ALL}")
+    print("-" * 40)
+    
+    # Without stealth
+    regular_antibot = await test_anti_bot_site(use_stealth=False)
+    print(f"{Fore.RED}Without Stealth:{Style.RESET_ALL}")
+    print(f"  • Success: {regular_antibot['success']}")
+    print(f"  • Blocked: {regular_antibot['blocked']}")
+    print(f"  • Status Code: {regular_antibot['status_code']}")
+    print(f"  • Page Title: {regular_antibot['page_title']}")
+    
+    # With stealth
+    stealth_antibot = await test_anti_bot_site(use_stealth=True)
+    print(f"\n{Fore.GREEN}With Stealth:{Style.RESET_ALL}")
+    print(f"  • Success: {stealth_antibot['success']}")
+    print(f"  • Blocked: {stealth_antibot['blocked']}")
+    print(f"  • Status Code: {stealth_antibot['status_code']}")
+    print(f"  • Page Title: {stealth_antibot['page_title']}")
+    
+    # Summary
+    print(f"\n{Fore.CYAN}{'='*60}{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}Summary:{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}{'='*60}{Style.RESET_ALL}")
+    print(f"\nStealth mode helps bypass basic bot detection by:")
+    print(f"  • Hiding webdriver property")
+    print(f"  • Modifying browser fingerprints")
+    print(f"  • Adjusting navigator properties")
+    print(f"  • Emulating real browser plugin behavior")
+    print(f"\n{Fore.YELLOW}Note:{Style.RESET_ALL} Stealth mode is not a silver bullet.")
+    print(f"Advanced anti-bot systems may still detect automation.")
+    print(f"Always respect robots.txt and website terms of service.")
+
+
+async def stealth_best_practices():
+    """Demonstrate best practices for using stealth mode"""
+    
+    print(f"\n\n{Fore.CYAN}{'='*60}{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}Stealth Mode Best Practices{Style.RESET_ALL}")
+    print(f"{Fore.CYAN}{'='*60}{Style.RESET_ALL}\n")
+    
+    # Best Practice 1: Combine with realistic behavior
+    print(f"{Fore.YELLOW}1. Combine with Realistic Behavior:{Style.RESET_ALL}")
+    
+    browser_config = BrowserConfig(
+        headless=False,
+        enable_stealth=True,
+        viewport_width=1920,
+        viewport_height=1080
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        # Simulate human-like behavior
+        human_behavior_script = """
+        (async () => {
+            // Wait random time between actions
+            const randomWait = () => Math.random() * 2000 + 1000;
+            
+            // Simulate reading
+            await new Promise(resolve => setTimeout(resolve, randomWait()));
+            
+            // Smooth scroll
+            const smoothScroll = async () => {
+                const totalHeight = document.body.scrollHeight;
+                const viewHeight = window.innerHeight;
+                let currentPosition = 0;
+                
+                while (currentPosition < totalHeight - viewHeight) {
+                    const scrollAmount = Math.random() * 300 + 100;
+                    window.scrollBy({
+                        top: scrollAmount,
+                        behavior: 'smooth'
+                    });
+                    currentPosition += scrollAmount;
+                    await new Promise(resolve => setTimeout(resolve, randomWait()));
+                }
+            };
+            
+            await smoothScroll();
+            console.log('Human-like behavior simulation completed');
+            return true;
+        })()
+        """
+        
+        config = CrawlerRunConfig(
+            js_code=human_behavior_script,
+            wait_until="networkidle",
+            delay_before_return_html=3.0,
+            capture_console_messages=True
+        )
+        
+        result = await crawler.arun(
+            url="https://example.com",
+            config=config
+        )
+        
+        print(f"  ✓ Simulated human-like scrolling and reading patterns")
+        print(f"  ✓ Added random delays between actions")
+        print(f"  ✓ Result: {result.success}")
+    
+    # Best Practice 2: Use appropriate viewport and user agent
+    print(f"\n{Fore.YELLOW}2. Use Realistic Viewport and User Agent:{Style.RESET_ALL}")
+    
+    # Get a realistic user agent
+    from crawl4ai.user_agent_generator import UserAgentGenerator
+    ua_generator = UserAgentGenerator()
+    
+    browser_config = BrowserConfig(
+        headless=True,
+        enable_stealth=True,
+        viewport_width=1920,
+        viewport_height=1080,
+        user_agent=ua_generator.generate(device_type="desktop", browser_type="chrome")
+    )
+    
+    print(f"  ✓ Using realistic viewport: 1920x1080")
+    print(f"  ✓ Using current Chrome user agent")
+    print(f"  ✓ Stealth mode will ensure consistency")
+    
+    # Best Practice 3: Manage request rate
+    print(f"\n{Fore.YELLOW}3. Manage Request Rate:{Style.RESET_ALL}")
+    print(f"  ✓ Add delays between requests")
+    print(f"  ✓ Randomize timing patterns")
+    print(f"  ✓ Respect robots.txt")
+    
+    # Best Practice 4: Session management
+    print(f"\n{Fore.YELLOW}4. Use Session Management:{Style.RESET_ALL}")
+    
+    browser_config = BrowserConfig(
+        headless=False,
+        enable_stealth=True
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        # Create a session for multiple requests
+        session_id = "stealth_session_1"
+        
+        config = CrawlerRunConfig(
+            session_id=session_id,
+            wait_until="domcontentloaded"
+        )
+        
+        # First request
+        result1 = await crawler.arun(
+            url="https://example.com",
+            config=config
+        )
+        
+        # Subsequent request reuses the same browser context
+        result2 = await crawler.arun(
+            url="https://example.com/about",
+            config=config
+        )
+        
+        print(f"  ✓ Reused browser session for multiple requests")
+        print(f"  ✓ Maintains cookies and state between requests")
+        print(f"  ✓ More efficient and realistic browsing pattern")
+    
+    print(f"\n{Fore.CYAN}{'='*60}{Style.RESET_ALL}")
+
+
+async def main():
+    """Run all examples"""
+    
+    # Run comparison tests
+    await compare_results()
+    
+    # Show best practices
+    await stealth_best_practices()
+    
+    print(f"\n{Fore.GREEN}Examples completed!{Style.RESET_ALL}")
+    print(f"\n{Fore.YELLOW}Remember:{Style.RESET_ALL}")
+    print(f"• Stealth mode helps with basic bot detection")
+    print(f"• Always respect website terms of service")
+    print(f"• Consider rate limiting and ethical scraping practices")
+    print(f"• For advanced protection, consider additional measures")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/stealth_mode_quick_start.py

@@ -0,0 +1,215 @@
+"""
+Quick Start: Using Stealth Mode in Crawl4AI
+
+This example shows practical use cases for the stealth mode feature.
+Stealth mode helps bypass basic bot detection mechanisms.
+"""
+
+import asyncio
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
+
+
+async def example_1_basic_stealth():
+    """Example 1: Basic stealth mode usage"""
+    print("\n=== Example 1: Basic Stealth Mode ===")
+    
+    # Enable stealth mode in browser config
+    browser_config = BrowserConfig(
+        enable_stealth=True,  # This is the key parameter
+        headless=True
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun(url="https://example.com")
+        print(f"✓ Crawled {result.url} successfully")
+        print(f"✓ Title: {result.metadata.get('title', 'N/A')}")
+
+
+async def example_2_stealth_with_screenshot():
+    """Example 2: Stealth mode with screenshot to show detection results"""
+    print("\n=== Example 2: Stealth Mode Visual Verification ===")
+    
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False  # Set to False to see the browser
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        config = CrawlerRunConfig(
+            screenshot=True,
+            wait_until="networkidle"
+        )
+        
+        result = await crawler.arun(
+            url="https://bot.sannysoft.com",
+            config=config
+        )
+        
+        if result.success:
+            print(f"✓ Successfully crawled bot detection site")
+            print(f"✓ With stealth enabled, many detection tests should show as passed")
+            
+            if result.screenshot:
+                # Save screenshot for verification
+                import base64
+                with open("stealth_detection_results.png", "wb") as f:
+                    f.write(base64.b64decode(result.screenshot))
+                print(f"✓ Screenshot saved as 'stealth_detection_results.png'")
+                print(f"  Check the screenshot to see detection results!")
+
+
+async def example_3_stealth_for_protected_sites():
+    """Example 3: Using stealth for sites with bot protection"""
+    print("\n=== Example 3: Stealth for Protected Sites ===")
+    
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=True,
+        viewport_width=1920,
+        viewport_height=1080
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        # Add human-like behavior
+        config = CrawlerRunConfig(
+            wait_until="networkidle",
+            delay_before_return_html=2.0,  # Wait 2 seconds
+            js_code="""
+            // Simulate human-like scrolling
+            window.scrollTo({
+                top: document.body.scrollHeight / 2,
+                behavior: 'smooth'
+            });
+            """
+        )
+        
+        # Try accessing a site that might have bot protection
+        result = await crawler.arun(
+            url="https://www.g2.com/products/slack/reviews",
+            config=config
+        )
+        
+        if result.success:
+            print(f"✓ Successfully accessed protected site")
+            print(f"✓ Retrieved {len(result.html)} characters of HTML")
+        else:
+            print(f"✗ Failed to access site: {result.error_message}")
+
+
+async def example_4_stealth_with_sessions():
+    """Example 4: Stealth mode with session management"""
+    print("\n=== Example 4: Stealth + Session Management ===")
+    
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        session_id = "my_stealth_session"
+        
+        # First request - establish session
+        config = CrawlerRunConfig(
+            session_id=session_id,
+            wait_until="domcontentloaded"
+        )
+        
+        result1 = await crawler.arun(
+            url="https://news.ycombinator.com",
+            config=config
+        )
+        print(f"✓ First request completed: {result1.url}")
+        
+        # Second request - reuse session
+        await asyncio.sleep(2)  # Brief delay between requests
+        
+        result2 = await crawler.arun(
+            url="https://news.ycombinator.com/best",
+            config=config
+        )
+        print(f"✓ Second request completed: {result2.url}")
+        print(f"✓ Session reused, maintaining cookies and state")
+
+
+async def example_5_stealth_comparison():
+    """Example 5: Compare results with and without stealth using screenshots"""
+    print("\n=== Example 5: Stealth Mode Comparison ===")
+    
+    test_url = "https://bot.sannysoft.com"
+    
+    # First test WITHOUT stealth
+    print("\nWithout stealth:")
+    regular_config = BrowserConfig(
+        enable_stealth=False,
+        headless=True
+    )
+    
+    async with AsyncWebCrawler(config=regular_config) as crawler:
+        config = CrawlerRunConfig(
+            screenshot=True,
+            wait_until="networkidle"
+        )
+        result = await crawler.arun(url=test_url, config=config)
+        
+        if result.success and result.screenshot:
+            import base64
+            with open("comparison_without_stealth.png", "wb") as f:
+                f.write(base64.b64decode(result.screenshot))
+            print(f"  ✓ Screenshot saved: comparison_without_stealth.png")
+            print(f"  Many tests will show as FAILED (red)")
+    
+    # Then test WITH stealth
+    print("\nWith stealth:")
+    stealth_config = BrowserConfig(
+        enable_stealth=True,
+        headless=True
+    )
+    
+    async with AsyncWebCrawler(config=stealth_config) as crawler:
+        config = CrawlerRunConfig(
+            screenshot=True,
+            wait_until="networkidle"
+        )
+        result = await crawler.arun(url=test_url, config=config)
+        
+        if result.success and result.screenshot:
+            import base64
+            with open("comparison_with_stealth.png", "wb") as f:
+                f.write(base64.b64decode(result.screenshot))
+            print(f"  ✓ Screenshot saved: comparison_with_stealth.png")
+            print(f"  More tests should show as PASSED (green)")
+    
+    print("\nCompare the two screenshots to see the difference!")
+
+
+async def main():
+    """Run all examples"""
+    print("Crawl4AI Stealth Mode Examples")
+    print("==============================")
+    
+    # Run basic example
+    await example_1_basic_stealth()
+    
+    # Run screenshot verification example
+    await example_2_stealth_with_screenshot()
+    
+    # Run protected site example
+    await example_3_stealth_for_protected_sites()
+    
+    # Run session example
+    await example_4_stealth_with_sessions()
+    
+    # Run comparison example
+    await example_5_stealth_comparison()
+    
+    print("\n" + "="*50)
+    print("Tips for using stealth mode effectively:")
+    print("- Use realistic viewport sizes (1920x1080, 1366x768)")
+    print("- Add delays between requests to appear more human")
+    print("- Combine with session management for better results")
+    print("- Remember: stealth mode is for legitimate scraping only")
+    print("="*50)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/stealth_test_simple.py

@@ -0,0 +1,62 @@
+"""
+Simple test to verify stealth mode is working
+"""
+
+import asyncio
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
+
+
+async def test_stealth():
+    """Test stealth mode effectiveness"""
+    
+    # Test WITHOUT stealth
+    print("=== WITHOUT Stealth ===")
+    config1 = BrowserConfig(
+        headless=False,
+        enable_stealth=False
+    )
+    
+    async with AsyncWebCrawler(config=config1) as crawler:
+        result = await crawler.arun(
+            url="https://bot.sannysoft.com",
+            config=CrawlerRunConfig(
+                wait_until="networkidle",
+                screenshot=True
+            )
+        )
+        print(f"Success: {result.success}")
+        # Take screenshot
+        if result.screenshot:
+            with open("without_stealth.png", "wb") as f:
+                import base64
+                f.write(base64.b64decode(result.screenshot))
+            print("Screenshot saved: without_stealth.png")
+    
+    # Test WITH stealth
+    print("\n=== WITH Stealth ===")
+    config2 = BrowserConfig(
+        headless=False,
+        enable_stealth=True
+    )
+    
+    async with AsyncWebCrawler(config=config2) as crawler:
+        result = await crawler.arun(
+            url="https://bot.sannysoft.com",
+            config=CrawlerRunConfig(
+                wait_until="networkidle",
+                screenshot=True
+            )
+        )
+        print(f"Success: {result.success}")
+        # Take screenshot
+        if result.screenshot:
+            with open("with_stealth.png", "wb") as f:
+                import base64
+                f.write(base64.b64decode(result.screenshot))
+            print("Screenshot saved: with_stealth.png")
+    
+    print("\nCheck the screenshots to see the difference in bot detection results!")
+
+
+if __name__ == "__main__":
+    asyncio.run(test_stealth())

docs/examples/undetectability/undetected_basic_test.py

@@ -0,0 +1,74 @@
+"""
+Basic Undetected Browser Test
+Simple example to test if undetected mode works
+"""
+
+import asyncio
+from crawl4ai import AsyncWebCrawler, BrowserConfig
+
+async def test_regular_mode():
+    """Test with regular browser"""
+    print("Testing Regular Browser Mode...")
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun(url="https://www.example.com")
+        print(f"Regular Mode - Success: {result.success}")
+        print(f"Regular Mode - Status: {result.status_code}")
+        print(f"Regular Mode - Content length: {len(result.markdown.raw_markdown)}")
+        print(f"Regular Mode - First 100 chars: {result.markdown.raw_markdown[:100]}...")
+        return result.success
+
+async def test_undetected_mode():
+    """Test with undetected browser"""
+    print("\nTesting Undetected Browser Mode...")
+    from crawl4ai import UndetectedAdapter
+    from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+    
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True
+    )
+    
+    # Create undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    
+    # Create strategy with undetected adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        result = await crawler.arun(url="https://www.example.com")
+        print(f"Undetected Mode - Success: {result.success}")
+        print(f"Undetected Mode - Status: {result.status_code}")
+        print(f"Undetected Mode - Content length: {len(result.markdown.raw_markdown)}")
+        print(f"Undetected Mode - First 100 chars: {result.markdown.raw_markdown[:100]}...")
+        return result.success
+
+async def main():
+    """Run both tests"""
+    print("🤖 Crawl4AI Basic Adapter Test\n")
+    
+    # Test regular mode
+    regular_success = await test_regular_mode()
+    
+    # Test undetected mode
+    undetected_success = await test_undetected_mode()
+    
+    # Summary
+    print("\n" + "="*50)
+    print("Summary:")
+    print(f"Regular Mode: {'✅ Success' if regular_success else '❌ Failed'}")
+    print(f"Undetected Mode: {'✅ Success' if undetected_success else '❌ Failed'}")
+    print("="*50)
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/undetectability/undetected_bot_test.py

@@ -0,0 +1,155 @@
+"""
+Bot Detection Test - Compare Regular vs Undetected
+Tests browser fingerprinting differences at bot.sannysoft.com
+"""
+
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler, 
+    BrowserConfig, 
+    CrawlerRunConfig,
+    UndetectedAdapter,
+    CrawlResult
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+# Bot detection test site
+TEST_URL = "https://bot.sannysoft.com"
+
+def analyze_bot_detection(result: CrawlResult) -> dict:
+    """Analyze bot detection results from the page"""
+    detections = {
+        "webdriver": False,
+        "headless": False, 
+        "automation": False,
+        "user_agent": False,
+        "total_tests": 0,
+        "failed_tests": 0
+    }
+    
+    if not result.success or not result.html:
+        return detections
+    
+    # Look for specific test results in the HTML
+    html_lower = result.html.lower()
+    
+    # Check for common bot indicators
+    if "webdriver" in html_lower and ("fail" in html_lower or "true" in html_lower):
+        detections["webdriver"] = True
+        detections["failed_tests"] += 1
+    
+    if "headless" in html_lower and ("fail" in html_lower or "true" in html_lower):
+        detections["headless"] = True
+        detections["failed_tests"] += 1
+    
+    if "automation" in html_lower and "detected" in html_lower:
+        detections["automation"] = True
+        detections["failed_tests"] += 1
+    
+    # Count total tests (approximate)
+    detections["total_tests"] = html_lower.count("test") + html_lower.count("check")
+    
+    return detections
+
+async def test_browser_mode(adapter_name: str, adapter=None):
+    """Test a browser mode and return results"""
+    print(f"\n{'='*60}")
+    print(f"Testing: {adapter_name}")
+    print(f"{'='*60}")
+    
+    browser_config = BrowserConfig(
+        headless=False,  # Run in headed mode for better results
+        verbose=True,
+        viewport_width=1920,
+        viewport_height=1080,
+    )
+    
+    if adapter:
+        # Use undetected mode
+        crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+            browser_config=browser_config,
+            browser_adapter=adapter
+        )
+        crawler = AsyncWebCrawler(
+            crawler_strategy=crawler_strategy,
+            config=browser_config
+        )
+    else:
+        # Use regular mode
+        crawler = AsyncWebCrawler(config=browser_config)
+    
+    async with crawler:
+        config = CrawlerRunConfig(
+            delay_before_return_html=3.0,  # Let detection scripts run
+            wait_for_images=True,
+            screenshot=True,
+            simulate_user=False,  # Don't simulate for accurate detection
+        )
+        
+        result = await crawler.arun(url=TEST_URL, config=config)
+        
+        print(f"\n✓ Success: {result.success}")
+        print(f"✓ Status Code: {result.status_code}")
+        
+        if result.success:
+            # Analyze detection results
+            detections = analyze_bot_detection(result)
+            
+            print(f"\n🔍 Bot Detection Analysis:")
+            print(f"  - WebDriver Detected: {'❌ Yes' if detections['webdriver'] else '✅ No'}")
+            print(f"  - Headless Detected: {'❌ Yes' if detections['headless'] else '✅ No'}")
+            print(f"  - Automation Detected: {'❌ Yes' if detections['automation'] else '✅ No'}")
+            print(f"  - Failed Tests: {detections['failed_tests']}")
+            
+            # Show some content
+            if result.markdown.raw_markdown:
+                print(f"\nContent preview:")
+                lines = result.markdown.raw_markdown.split('\n')
+                for line in lines[:20]:  # Show first 20 lines
+                    if any(keyword in line.lower() for keyword in ['test', 'pass', 'fail', 'yes', 'no']):
+                        print(f"  {line.strip()}")
+        
+        return result, detections if result.success else {}
+
+async def main():
+    """Run the comparison"""
+    print("🤖 Crawl4AI - Bot Detection Test")
+    print(f"Testing at: {TEST_URL}")
+    print("This site runs various browser fingerprinting tests\n")
+    
+    # Test regular browser
+    regular_result, regular_detections = await test_browser_mode("Regular Browser")
+    
+    # Small delay
+    await asyncio.sleep(2)
+    
+    # Test undetected browser
+    undetected_adapter = UndetectedAdapter()
+    undetected_result, undetected_detections = await test_browser_mode(
+        "Undetected Browser", 
+        undetected_adapter
+    )
+    
+    # Summary comparison
+    print(f"\n{'='*60}")
+    print("COMPARISON SUMMARY")
+    print(f"{'='*60}")
+    
+    print(f"\n{'Test':<25} {'Regular':<15} {'Undetected':<15}")
+    print(f"{'-'*55}")
+    
+    if regular_detections and undetected_detections:
+        print(f"{'WebDriver Detection':<25} {'❌ Detected' if regular_detections['webdriver'] else '✅ Passed':<15} {'❌ Detected' if undetected_detections['webdriver'] else '✅ Passed':<15}")
+        print(f"{'Headless Detection':<25} {'❌ Detected' if regular_detections['headless'] else '✅ Passed':<15} {'❌ Detected' if undetected_detections['headless'] else '✅ Passed':<15}")
+        print(f"{'Automation Detection':<25} {'❌ Detected' if regular_detections['automation'] else '✅ Passed':<15} {'❌ Detected' if undetected_detections['automation'] else '✅ Passed':<15}")
+        print(f"{'Failed Tests':<25} {regular_detections['failed_tests']:<15} {undetected_detections['failed_tests']:<15}")
+    
+    print(f"\n{'='*60}")
+    
+    if undetected_detections.get('failed_tests', 0) < regular_detections.get('failed_tests', 1):
+        print("✅ Undetected browser performed better at evading detection!")
+    else:
+        print("ℹ️  Both browsers had similar detection results")
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/undetectability/undetected_cloudflare_test.py

@@ -0,0 +1,164 @@
+"""
+Undetected Browser Test - Cloudflare Protected Site
+Tests the difference between regular and undetected modes on a Cloudflare-protected site
+"""
+
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler, 
+    BrowserConfig, 
+    CrawlerRunConfig,
+    UndetectedAdapter
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+# Test URL with Cloudflare protection
+TEST_URL = "https://nowsecure.nl"
+
+async def test_regular_browser():
+    """Test with regular browser - likely to be blocked"""
+    print("=" * 60)
+    print("Testing with Regular Browser")
+    print("=" * 60)
+    
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True,
+        viewport_width=1920,
+        viewport_height=1080,
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        config = CrawlerRunConfig(
+            delay_before_return_html=2.0,
+            simulate_user=True,
+            magic=True,  # Try with magic mode too
+        )
+        
+        result = await crawler.arun(url=TEST_URL, config=config)
+        
+        print(f"\n✓ Success: {result.success}")
+        print(f"✓ Status Code: {result.status_code}")
+        print(f"✓ HTML Length: {len(result.html)}")
+        
+        # Check for Cloudflare challenge
+        if result.html:
+            cf_indicators = [
+                "Checking your browser",
+                "Please stand by",
+                "cloudflare",
+                "cf-browser-verification",
+                "Access denied",
+                "Ray ID"
+            ]
+            
+            detected = False
+            for indicator in cf_indicators:
+                if indicator.lower() in result.html.lower():
+                    print(f"⚠️  Cloudflare Challenge Detected: '{indicator}' found")
+                    detected = True
+                    break
+            
+            if not detected and len(result.markdown.raw_markdown) > 100:
+                print("✅ Successfully bypassed Cloudflare!")
+                print(f"Content preview: {result.markdown.raw_markdown[:200]}...")
+            elif not detected:
+                print("⚠️  Page loaded but content seems minimal")
+        
+        return result
+
+async def test_undetected_browser():
+    """Test with undetected browser - should bypass Cloudflare"""
+    print("\n" + "=" * 60)
+    print("Testing with Undetected Browser")
+    print("=" * 60)
+    
+    browser_config = BrowserConfig(
+        headless=False,  # Headless is easier to detect
+        verbose=True,
+        viewport_width=1920,
+        viewport_height=1080,
+    )
+    
+    # Create undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    
+    # Create strategy with undetected adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        config = CrawlerRunConfig(
+            delay_before_return_html=2.0,
+            simulate_user=True,
+        )
+        
+        result = await crawler.arun(url=TEST_URL, config=config)
+        
+        print(f"\n✓ Success: {result.success}")
+        print(f"✓ Status Code: {result.status_code}")
+        print(f"✓ HTML Length: {len(result.html)}")
+        
+        # Check for Cloudflare challenge
+        if result.html:
+            cf_indicators = [
+                "Checking your browser",
+                "Please stand by",
+                "cloudflare",
+                "cf-browser-verification",
+                "Access denied",
+                "Ray ID"
+            ]
+            
+            detected = False
+            for indicator in cf_indicators:
+                if indicator.lower() in result.html.lower():
+                    print(f"⚠️  Cloudflare Challenge Detected: '{indicator}' found")
+                    detected = True
+                    break
+            
+            if not detected and len(result.markdown.raw_markdown) > 100:
+                print("✅ Successfully bypassed Cloudflare!")
+                print(f"Content preview: {result.markdown.raw_markdown[:200]}...")
+            elif not detected:
+                print("⚠️  Page loaded but content seems minimal")
+        
+        return result
+
+async def main():
+    """Compare regular vs undetected browser"""
+    print("🤖 Crawl4AI - Cloudflare Bypass Test")
+    print(f"Testing URL: {TEST_URL}\n")
+    
+    # Test regular browser
+    regular_result = await test_regular_browser()
+    
+    # Small delay
+    await asyncio.sleep(2)
+    
+    # Test undetected browser
+    undetected_result = await test_undetected_browser()
+    
+    # Summary
+    print("\n" + "=" * 60)
+    print("SUMMARY")
+    print("=" * 60)
+    print(f"Regular Browser:")
+    print(f"  - Success: {regular_result.success}")
+    print(f"  - Content Length: {len(regular_result.markdown.raw_markdown) if regular_result.markdown else 0}")
+    
+    print(f"\nUndetected Browser:")
+    print(f"  - Success: {undetected_result.success}")
+    print(f"  - Content Length: {len(undetected_result.markdown.raw_markdown) if undetected_result.markdown else 0}")
+    
+    if undetected_result.success and len(undetected_result.markdown.raw_markdown) > len(regular_result.markdown.raw_markdown):
+        print("\n✅ Undetected browser successfully bypassed protection!")
+    print("=" * 60)
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/examples/undetectability/undetected_vs_regular_comparison.py

@@ -0,0 +1,184 @@
+"""
+Undetected vs Regular Browser Comparison
+This example demonstrates the difference between regular and undetected browser modes
+when accessing sites with bot detection services.
+
+Based on tested anti-bot services:
+- Cloudflare
+- Kasada
+- Akamai
+- DataDome
+- Bet365
+- And others
+"""
+
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler,
+    BrowserConfig,
+    CrawlerRunConfig,
+    PlaywrightAdapter,
+    UndetectedAdapter,
+    CrawlResult
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+
+# Test URLs for various bot detection services
+TEST_SITES = {
+    "Cloudflare Protected": "https://nowsecure.nl",
+    # "Bot Detection Test": "https://bot.sannysoft.com",
+    # "Fingerprint Test": "https://fingerprint.com/products/bot-detection",
+    # "Browser Scan": "https://browserscan.net",
+    # "CreepJS": "https://abrahamjuliot.github.io/creepjs",
+}
+
+
+async def test_with_adapter(url: str, adapter_name: str, adapter):
+    """Test a URL with a specific adapter"""
+    browser_config = BrowserConfig(
+        headless=False,  # Better for avoiding detection
+        viewport_width=1920,
+        viewport_height=1080,
+        verbose=True,
+    )
+    
+    # Create the crawler strategy with the adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=adapter
+    )
+    
+    print(f"\n{'='*60}")
+    print(f"Testing with {adapter_name} adapter")
+    print(f"URL: {url}")
+    print(f"{'='*60}")
+    
+    try:
+        async with AsyncWebCrawler(
+            crawler_strategy=crawler_strategy,
+            config=browser_config
+        ) as crawler:
+            crawler_config = CrawlerRunConfig(
+                delay_before_return_html=3.0,  # Give page time to load
+                wait_for_images=True,
+                screenshot=True,
+                simulate_user=True,  # Add user simulation
+            )
+            
+            result: CrawlResult = await crawler.arun(
+                url=url,
+                config=crawler_config
+            )
+            
+            # Check results
+            print(f"✓ Status Code: {result.status_code}")
+            print(f"✓ Success: {result.success}")
+            print(f"✓ HTML Length: {len(result.html)}")
+            print(f"✓ Markdown Length: {len(result.markdown.raw_markdown)}")
+            
+            # Check for common bot detection indicators
+            detection_indicators = [
+                "Access denied",
+                "Please verify you are human",
+                "Checking your browser",
+                "Enable JavaScript",
+                "captcha",
+                "403 Forbidden",
+                "Bot detection",
+                "Security check"
+            ]
+            
+            content_lower = result.markdown.raw_markdown.lower()
+            detected = False
+            for indicator in detection_indicators:
+                if indicator.lower() in content_lower:
+                    print(f"⚠️  Possible detection: Found '{indicator}'")
+                    detected = True
+                    break
+            
+            if not detected:
+                print("✅ No obvious bot detection triggered!")
+                # Show first 200 chars of content
+                print(f"Content preview: {result.markdown.raw_markdown[:200]}...")
+            
+            return result.success and not detected
+            
+    except Exception as e:
+        print(f"❌ Error: {str(e)}")
+        return False
+
+
+async def compare_adapters(url: str, site_name: str):
+    """Compare regular and undetected adapters on the same URL"""
+    print(f"\n{'#'*60}")
+    print(f"# Testing: {site_name}")
+    print(f"{'#'*60}")
+    
+    # Test with regular adapter
+    regular_adapter = PlaywrightAdapter()
+    regular_success = await test_with_adapter(url, "Regular", regular_adapter)
+    
+    # Small delay between tests
+    await asyncio.sleep(2)
+    
+    # Test with undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    undetected_success = await test_with_adapter(url, "Undetected", undetected_adapter)
+    
+    # Summary
+    print(f"\n{'='*60}")
+    print(f"Summary for {site_name}:")
+    print(f"Regular Adapter: {'✅ Passed' if regular_success else '❌ Blocked/Detected'}")
+    print(f"Undetected Adapter: {'✅ Passed' if undetected_success else '❌ Blocked/Detected'}")
+    print(f"{'='*60}")
+    
+    return regular_success, undetected_success
+
+
+async def main():
+    """Run comparison tests on multiple sites"""
+    print("🤖 Crawl4AI Browser Adapter Comparison")
+    print("Testing regular vs undetected browser modes\n")
+    
+    results = {}
+    
+    # Test each site
+    for site_name, url in TEST_SITES.items():
+        regular, undetected = await compare_adapters(url, site_name)
+        results[site_name] = {
+            "regular": regular,
+            "undetected": undetected
+        }
+        
+        # Delay between different sites
+        await asyncio.sleep(3)
+    
+    # Final summary
+    print(f"\n{'#'*60}")
+    print("# FINAL RESULTS")
+    print(f"{'#'*60}")
+    print(f"{'Site':<30} {'Regular':<15} {'Undetected':<15}")
+    print(f"{'-'*60}")
+    
+    for site, result in results.items():
+        regular_status = "✅ Passed" if result["regular"] else "❌ Blocked"
+        undetected_status = "✅ Passed" if result["undetected"] else "❌ Blocked"
+        print(f"{site:<30} {regular_status:<15} {undetected_status:<15}")
+    
+    # Calculate success rates
+    regular_success = sum(1 for r in results.values() if r["regular"])
+    undetected_success = sum(1 for r in results.values() if r["undetected"])
+    total = len(results)
+    
+    print(f"\n{'='*60}")
+    print(f"Success Rates:")
+    print(f"Regular Adapter: {regular_success}/{total} ({regular_success/total*100:.1f}%)")
+    print(f"Undetected Adapter: {undetected_success}/{total} ({undetected_success/total*100:.1f}%)")
+    print(f"{'='*60}")
+
+
+if __name__ == "__main__":
+    # Note: This example may take a while to run as it tests multiple sites
+    # You can comment out sites in TEST_SITES to run faster tests
+    asyncio.run(main())

docs/examples/undetected_simple_demo.py

@@ -0,0 +1,118 @@
+"""
+Simple Undetected Browser Demo
+Demonstrates the basic usage of undetected browser mode
+"""
+
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler, 
+    BrowserConfig, 
+    CrawlerRunConfig,
+    UndetectedAdapter
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+async def crawl_with_regular_browser(url: str):
+    """Crawl with regular browser"""
+    print("\n[Regular Browser Mode]")
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True,
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun(
+            url=url,
+            config=CrawlerRunConfig(
+                delay_before_return_html=2.0
+            )
+        )
+        
+        print(f"Success: {result.success}")
+        print(f"Status: {result.status_code}")
+        print(f"Content length: {len(result.markdown.raw_markdown)}")
+        
+        # Check for bot detection keywords
+        content = result.markdown.raw_markdown.lower()
+        if any(word in content for word in ["cloudflare", "checking your browser", "please wait"]):
+            print("⚠️  Bot detection triggered!")
+        else:
+            print("✅ Page loaded successfully")
+        
+        return result
+
+async def crawl_with_undetected_browser(url: str):
+    """Crawl with undetected browser"""
+    print("\n[Undetected Browser Mode]")
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True,
+    )
+    
+    # Create undetected adapter and strategy
+    undetected_adapter = UndetectedAdapter()
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        result = await crawler.arun(
+            url=url,
+            config=CrawlerRunConfig(
+                delay_before_return_html=2.0
+            )
+        )
+        
+        print(f"Success: {result.success}")
+        print(f"Status: {result.status_code}")
+        print(f"Content length: {len(result.markdown.raw_markdown)}")
+        
+        # Check for bot detection keywords
+        content = result.markdown.raw_markdown.lower()
+        if any(word in content for word in ["cloudflare", "checking your browser", "please wait"]):
+            print("⚠️  Bot detection triggered!")
+        else:
+            print("✅ Page loaded successfully")
+        
+        return result
+
+async def main():
+    """Demo comparing regular vs undetected modes"""
+    print("🤖 Crawl4AI Undetected Browser Demo")
+    print("="*50)
+    
+    # Test URLs - you can change these
+    test_urls = [
+        "https://www.example.com",  # Simple site
+        "https://httpbin.org/headers",  # Shows request headers
+    ]
+    
+    for url in test_urls:
+        print(f"\n📍 Testing URL: {url}")
+        
+        # Test with regular browser
+        regular_result = await crawl_with_regular_browser(url)
+        
+        # Small delay
+        await asyncio.sleep(2)
+        
+        # Test with undetected browser
+        undetected_result = await crawl_with_undetected_browser(url)
+        
+        # Compare results
+        print(f"\n📊 Comparison for {url}:")
+        print(f"Regular browser content: {len(regular_result.markdown.raw_markdown)} chars")
+        print(f"Undetected browser content: {len(undetected_result.markdown.raw_markdown)} chars")
+        
+        if url == "https://httpbin.org/headers":
+            # Show headers for comparison
+            print("\nHeaders seen by server:")
+            print("Regular:", regular_result.markdown.raw_markdown[:500])
+            print("\nUndetected:", undetected_result.markdown.raw_markdown[:500])
+
+if __name__ == "__main__":
+    asyncio.run(main())

docs/md_v2/advanced/adaptive-strategies.md

@@ -0,0 +1,432 @@
+# Advanced Adaptive Strategies
+
+## Overview
+
+While the default adaptive crawling configuration works well for most use cases, understanding the underlying strategies and scoring mechanisms allows you to fine-tune the crawler for specific domains and requirements.
+
+## The Three-Layer Scoring System
+
+### 1. Coverage Score
+
+Coverage measures how comprehensively your knowledge base covers the query terms and related concepts.
+
+#### Mathematical Foundation
+
+```python
+Coverage(K, Q) = Σ(t ∈ Q) score(t, K) / |Q|
+
+where score(t, K) = doc_coverage(t) × (1 + freq_boost(t))
+```
+
+#### Components
+
+- **Document Coverage**: Percentage of documents containing the term
+- **Frequency Boost**: Logarithmic bonus for term frequency
+- **Query Decomposition**: Handles multi-word queries intelligently
+
+#### Tuning Coverage
+
+```python
+# For technical documentation with specific terminology
+config = AdaptiveConfig(
+    confidence_threshold=0.85,  # Require high coverage
+    top_k_links=5              # Cast wider net
+)
+
+# For general topics with synonyms
+config = AdaptiveConfig(
+    confidence_threshold=0.6,   # Lower threshold
+    top_k_links=2              # More focused
+)
+```
+
+### 2. Consistency Score
+
+Consistency evaluates whether the information across pages is coherent and non-contradictory.
+
+#### How It Works
+
+1. Extracts key statements from each document
+2. Compares statements across documents
+3. Measures agreement vs. contradiction
+4. Returns normalized score (0-1)
+
+#### Practical Impact
+
+- **High consistency (>0.8)**: Information is reliable and coherent
+- **Medium consistency (0.5-0.8)**: Some variation, but generally aligned
+- **Low consistency (<0.5)**: Conflicting information, need more sources
+
+### 3. Saturation Score
+
+Saturation detects when new pages stop providing novel information.
+
+#### Detection Algorithm
+
+```python
+# Tracks new unique terms per page
+new_terms_page_1 = 50
+new_terms_page_2 = 30  # 60% of first
+new_terms_page_3 = 15  # 50% of second
+new_terms_page_4 = 5   # 33% of third
+# Saturation detected: rapidly diminishing returns
+```
+
+#### Configuration
+
+```python
+config = AdaptiveConfig(
+    min_gain_threshold=0.1  # Stop if <10% new information
+)
+```
+
+## Link Ranking Algorithm
+
+### Expected Information Gain
+
+Each uncrawled link is scored based on:
+
+```python
+ExpectedGain(link) = Relevance × Novelty × Authority
+```
+
+#### 1. Relevance Scoring
+
+Uses BM25 algorithm on link preview text:
+
+```python
+relevance = BM25(link.preview_text, query)
+```
+
+Factors:
+- Term frequency in preview
+- Inverse document frequency
+- Preview length normalization
+
+#### 2. Novelty Estimation
+
+Measures how different the link appears from already-crawled content:
+
+```python
+novelty = 1 - max_similarity(preview, knowledge_base)
+```
+
+Prevents crawling duplicate or highly similar pages.
+
+#### 3. Authority Calculation
+
+URL structure and domain analysis:
+
+```python
+authority = f(domain_rank, url_depth, url_structure)
+```
+
+Factors:
+- Domain reputation
+- URL depth (fewer slashes = higher authority)
+- Clean URL structure
+
+### Custom Link Scoring
+
+```python
+class CustomLinkScorer:
+    def score(self, link: Link, query: str, state: CrawlState) -> float:
+        # Prioritize specific URL patterns
+        if "/api/reference/" in link.href:
+            return 2.0  # Double the score
+        
+        # Deprioritize certain sections
+        if "/archive/" in link.href:
+            return 0.1  # Reduce score by 90%
+        
+        # Default scoring
+        return 1.0
+
+# Use with adaptive crawler
+adaptive = AdaptiveCrawler(
+    crawler,
+    config=config,
+    link_scorer=CustomLinkScorer()
+)
+```
+
+## Domain-Specific Configurations
+
+### Technical Documentation
+
+```python
+tech_doc_config = AdaptiveConfig(
+    confidence_threshold=0.85,
+    max_pages=30,
+    top_k_links=3,
+    min_gain_threshold=0.05  # Keep crawling for small gains
+)
+```
+
+Rationale:
+- High threshold ensures comprehensive coverage
+- Lower gain threshold captures edge cases
+- Moderate link following for depth
+
+### News & Articles
+
+```python
+news_config = AdaptiveConfig(
+    confidence_threshold=0.6,
+    max_pages=10,
+    top_k_links=5,
+    min_gain_threshold=0.15  # Stop quickly on repetition
+)
+```
+
+Rationale:
+- Lower threshold (articles often repeat information)
+- Higher gain threshold (avoid duplicate stories)
+- More links per page (explore different perspectives)
+
+### E-commerce
+
+```python
+ecommerce_config = AdaptiveConfig(
+    confidence_threshold=0.7,
+    max_pages=20,
+    top_k_links=2,
+    min_gain_threshold=0.1
+)
+```
+
+Rationale:
+- Balanced threshold for product variations
+- Focused link following (avoid infinite products)
+- Standard gain threshold
+
+### Research & Academic
+
+```python
+research_config = AdaptiveConfig(
+    confidence_threshold=0.9,
+    max_pages=50,
+    top_k_links=4,
+    min_gain_threshold=0.02  # Very low - capture citations
+)
+```
+
+Rationale:
+- Very high threshold for completeness
+- Many pages allowed for thorough research
+- Very low gain threshold to capture references
+
+## Performance Optimization
+
+### Memory Management
+
+```python
+# For large crawls, use streaming
+config = AdaptiveConfig(
+    max_pages=100,
+    save_state=True,
+    state_path="large_crawl.json"
+)
+
+# Periodically clean state
+if len(state.knowledge_base) > 1000:
+    # Keep only most relevant
+    state.knowledge_base = get_top_relevant(state.knowledge_base, 500)
+```
+
+### Parallel Processing
+
+```python
+# Use multiple start points
+start_urls = [
+    "https://docs.example.com/intro",
+    "https://docs.example.com/api",
+    "https://docs.example.com/guides"
+]
+
+# Crawl in parallel
+tasks = [
+    adaptive.digest(url, query)
+    for url in start_urls
+]
+results = await asyncio.gather(*tasks)
+```
+
+### Caching Strategy
+
+```python
+# Enable caching for repeated crawls
+async with AsyncWebCrawler(
+    config=BrowserConfig(
+        cache_mode=CacheMode.ENABLED
+    )
+) as crawler:
+    adaptive = AdaptiveCrawler(crawler, config)
+```
+
+## Debugging & Analysis
+
+### Enable Verbose Logging
+
+```python
+import logging
+
+logging.basicConfig(level=logging.DEBUG)
+adaptive = AdaptiveCrawler(crawler, config, verbose=True)
+```
+
+### Analyze Crawl Patterns
+
+```python
+# After crawling
+state = await adaptive.digest(start_url, query)
+
+# Analyze link selection
+print("Link selection order:")
+for i, url in enumerate(state.crawl_order):
+    print(f"{i+1}. {url}")
+
+# Analyze term discovery
+print("\nTerm discovery rate:")
+for i, new_terms in enumerate(state.new_terms_history):
+    print(f"Page {i+1}: {new_terms} new terms")
+
+# Analyze score progression
+print("\nScore progression:")
+print(f"Coverage: {state.metrics['coverage_history']}")
+print(f"Saturation: {state.metrics['saturation_history']}")
+```
+
+### Export for Analysis
+
+```python
+# Export detailed metrics
+import json
+
+metrics = {
+    "query": query,
+    "total_pages": len(state.crawled_urls),
+    "confidence": adaptive.confidence,
+    "coverage_stats": adaptive.coverage_stats,
+    "crawl_order": state.crawl_order,
+    "term_frequencies": dict(state.term_frequencies),
+    "new_terms_history": state.new_terms_history
+}
+
+with open("crawl_analysis.json", "w") as f:
+    json.dump(metrics, f, indent=2)
+```
+
+## Custom Strategies
+
+### Implementing a Custom Strategy
+
+```python
+from crawl4ai.adaptive_crawler import BaseStrategy
+
+class DomainSpecificStrategy(BaseStrategy):
+    def calculate_coverage(self, state: CrawlState) -> float:
+        # Custom coverage calculation
+        # e.g., weight certain terms more heavily
+        pass
+    
+    def calculate_consistency(self, state: CrawlState) -> float:
+        # Custom consistency logic
+        # e.g., domain-specific validation
+        pass
+    
+    def rank_links(self, links: List[Link], state: CrawlState) -> List[Link]:
+        # Custom link ranking
+        # e.g., prioritize specific URL patterns
+        pass
+
+# Use custom strategy
+adaptive = AdaptiveCrawler(
+    crawler,
+    config=config,
+    strategy=DomainSpecificStrategy()
+)
+```
+
+### Combining Strategies
+
+```python
+class HybridStrategy(BaseStrategy):
+    def __init__(self):
+        self.strategies = [
+            TechnicalDocStrategy(),
+            SemanticSimilarityStrategy(),
+            URLPatternStrategy()
+        ]
+    
+    def calculate_confidence(self, state: CrawlState) -> float:
+        # Weighted combination of strategies
+        scores = [s.calculate_confidence(state) for s in self.strategies]
+        weights = [0.5, 0.3, 0.2]
+        return sum(s * w for s, w in zip(scores, weights))
+```
+
+## Best Practices
+
+### 1. Start Conservative
+
+Begin with default settings and adjust based on results:
+
+```python
+# Start with defaults
+result = await adaptive.digest(url, query)
+
+# Analyze and adjust
+if adaptive.confidence < 0.7:
+    config.max_pages += 10
+    config.confidence_threshold -= 0.1
+```
+
+### 2. Monitor Resource Usage
+
+```python
+import psutil
+
+# Check memory before large crawls
+memory_percent = psutil.virtual_memory().percent
+if memory_percent > 80:
+    config.max_pages = min(config.max_pages, 20)
+```
+
+### 3. Use Domain Knowledge
+
+```python
+# For API documentation
+if "api" in start_url:
+    config.top_k_links = 2  # APIs have clear structure
+    
+# For blogs
+if "blog" in start_url:
+    config.min_gain_threshold = 0.2  # Avoid similar posts
+```
+
+### 4. Validate Results
+
+```python
+# Always validate the knowledge base
+relevant_content = adaptive.get_relevant_content(top_k=10)
+
+# Check coverage
+query_terms = set(query.lower().split())
+covered_terms = set()
+
+for doc in relevant_content:
+    content_lower = doc['content'].lower()
+    for term in query_terms:
+        if term in content_lower:
+            covered_terms.add(term)
+
+coverage_ratio = len(covered_terms) / len(query_terms)
+print(f"Query term coverage: {coverage_ratio:.0%}")
+```
+
+## Next Steps
+
+- Explore [Custom Strategy Implementation](../tutorials/custom-adaptive-strategies.md)
+- Learn about [Knowledge Base Management](../tutorials/knowledge-base-management.md)
+- See [Performance Benchmarks](../benchmarks/adaptive-performance.md)

docs/md_v2/advanced/advanced-features.md

@@ -66,29 +66,38 @@ Sometimes you need a visual record of a page or a PDF “printout.” Crawl4AI c
 ```python
 import os, asyncio
 from base64 import b64decode
-from crawl4ai import AsyncWebCrawler, CacheMode
+from crawl4ai import AsyncWebCrawler, CacheMode, CrawlerRunConfig
 
 async def main():
+    run_config = CrawlerRunConfig(
+        cache_mode=CacheMode.BYPASS,
+        screenshot=True,
+        pdf=True
+    )
+
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(
             url="https://en.wikipedia.org/wiki/List_of_common_misconceptions",
-            cache_mode=CacheMode.BYPASS,
-            pdf=True,
-            screenshot=True
+            config=run_config
         )
-        
         if result.success:
-            # Save screenshot
+            print(f"Screenshot data present: {result.screenshot is not None}")
+            print(f"PDF data present: {result.pdf is not None}")
+
             if result.screenshot:
+                print(f"[OK] Screenshot captured, size: {len(result.screenshot)} bytes")
                 with open("wikipedia_screenshot.png", "wb") as f:
                     f.write(b64decode(result.screenshot))
-            
-            # Save PDF
+            else:
+                print("[WARN] Screenshot data is None.")
+
             if result.pdf:
+                print(f"[OK] PDF captured, size: {len(result.pdf)} bytes")
                 with open("wikipedia_page.pdf", "wb") as f:
                     f.write(result.pdf)
-            
-            print("[OK] PDF & screenshot captured.")
+            else:
+                print("[WARN] PDF data is None.")
+
         else:
             print("[ERROR]", result.error_message)
 
@@ -349,9 +358,77 @@ if __name__ == "__main__":
 
 ---
 
+---
+
+## 7. Anti-Bot Features (Stealth Mode & Undetected Browser)
+
+Crawl4AI provides two powerful features to bypass bot detection:
+
+### 7.1 Stealth Mode
+
+Stealth mode uses playwright-stealth to modify browser fingerprints and behaviors. Enable it with a simple flag:
+
+```python
+browser_config = BrowserConfig(
+    enable_stealth=True,  # Activates stealth mode
+    headless=False
+)
+```
+
+**When to use**: Sites with basic bot detection (checking navigator.webdriver, plugins, etc.)
+
+### 7.2 Undetected Browser
+
+For advanced bot detection, use the undetected browser adapter:
+
+```python
+from crawl4ai import UndetectedAdapter
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+# Create undetected adapter
+adapter = UndetectedAdapter()
+strategy = AsyncPlaywrightCrawlerStrategy(
+    browser_config=browser_config,
+    browser_adapter=adapter
+)
+
+async with AsyncWebCrawler(crawler_strategy=strategy, config=browser_config) as crawler:
+    # Your crawling code
+```
+
+**When to use**: Sites with sophisticated bot detection (Cloudflare, DataDome, etc.)
+
+### 7.3 Combining Both
+
+For maximum evasion, combine stealth mode with undetected browser:
+
+```python
+browser_config = BrowserConfig(
+    enable_stealth=True,  # Enable stealth
+    headless=False
+)
+
+adapter = UndetectedAdapter()  # Use undetected browser
+```
+
+### Choosing the Right Approach
+
+| Detection Level | Recommended Approach |
+|----------------|---------------------|
+| No protection | Regular browser |
+| Basic checks | Regular + Stealth mode |
+| Advanced protection | Undetected browser |
+| Maximum evasion | Undetected + Stealth mode |
+
+**Best Practice**: Start with regular browser + stealth mode. Only use undetected browser if needed, as it may be slightly slower.
+
+See [Undetected Browser Mode](undetected-browser.md) for detailed examples.
+
+---
+
 ## Conclusion & Next Steps
 
-You’ve now explored several **advanced** features:
+You've now explored several **advanced** features:
 
 - **Proxy Usage**  
 - **PDF & Screenshot** capturing for large or critical pages  
@@ -359,7 +436,10 @@ You’ve now explored several **advanced** features:
 - **Custom Headers** for language or specialized requests  
 - **Session Persistence** via storage state
 - **Robots.txt Compliance**
+- **Anti-Bot Features** (Stealth Mode & Undetected Browser)
 
-With these power tools, you can build robust scraping workflows that mimic real user behavior, handle secure sites, capture detailed snapshots, and manage sessions across multiple runs—streamlining your entire data collection pipeline.
+With these power tools, you can build robust scraping workflows that mimic real user behavior, handle secure sites, capture detailed snapshots, manage sessions across multiple runs, and bypass bot detection—streamlining your entire data collection pipeline.
 
-**Last Updated**: 2025-01-01
+**Note**: In future versions, we may enable stealth mode and undetected browser by default. For now, users should explicitly enable these features when needed.
+
+**Last Updated**: 2025-01-17

docs/md_v2/advanced/multi-url-crawling.md

@@ -404,7 +404,182 @@ for result in results:
         print(f"Duration: {dr.end_time - dr.start_time}")
 ```
 
-## 6. Summary
+## 6. URL-Specific Configurations
+
+When crawling diverse content types, you often need different configurations for different URLs. For example:
+- PDFs need specialized extraction
+- Blog pages benefit from content filtering
+- Dynamic sites need JavaScript execution
+- API endpoints need JSON parsing
+
+### 6.1 Basic URL Pattern Matching
+
+```python
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, MatchMode
+from crawl4ai.processors.pdf import PDFContentScrapingStrategy
+from crawl4ai.extraction_strategy import JsonCssExtractionStrategy
+from crawl4ai.content_filter_strategy import PruningContentFilter
+from crawl4ai.markdown_generation_strategy import DefaultMarkdownGenerator
+
+async def crawl_mixed_content():
+    # Configure different strategies for different content
+    configs = [
+        # PDF files - specialized extraction
+        CrawlerRunConfig(
+            url_matcher="*.pdf",
+            scraping_strategy=PDFContentScrapingStrategy()
+        ),
+        
+        # Blog/article pages - content filtering
+        CrawlerRunConfig(
+            url_matcher=["*/blog/*", "*/article/*"],
+            markdown_generator=DefaultMarkdownGenerator(
+                content_filter=PruningContentFilter(threshold=0.48)
+            )
+        ),
+        
+        # Dynamic pages - JavaScript execution
+        CrawlerRunConfig(
+            url_matcher=lambda url: 'github.com' in url,
+            js_code="window.scrollTo(0, 500);"
+        ),
+        
+        # API endpoints - JSON extraction
+        CrawlerRunConfig(
+            url_matcher=lambda url: 'api' in url or url.endswith('.json'),
+            # Custome settings for JSON extraction
+        ),
+        
+        # Default config for everything else
+        CrawlerRunConfig()  # No url_matcher means it matches ALL URLs (fallback)
+    ]
+    
+    # Mixed URLs
+    urls = [
+        "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf",
+        "https://blog.python.org/",
+        "https://github.com/microsoft/playwright",
+        "https://httpbin.org/json",
+        "https://example.com/"
+    ]
+    
+    async with AsyncWebCrawler() as crawler:
+        results = await crawler.arun_many(
+            urls=urls,
+            config=configs  # Pass list of configs
+        )
+        
+        for result in results:
+            print(f"{result.url}: {len(result.markdown)} chars")
+```
+
+### 6.2 Advanced Pattern Matching
+
+**Important**: A `CrawlerRunConfig` without `url_matcher` (or with `url_matcher=None`) matches ALL URLs. This makes it perfect as a default/fallback configuration.
+
+The `url_matcher` parameter supports three types of patterns:
+
+#### Glob Patterns (Strings)
+```python
+# Simple patterns
+"*.pdf"                    # Any PDF file
+"*/api/*"                  # Any URL with /api/ in path
+"https://*.example.com/*"  # Subdomain matching
+"*://example.com/blog/*"   # Any protocol
+```
+
+#### Custom Functions
+```python
+# Complex logic with lambdas
+lambda url: url.startswith('https://') and 'secure' in url
+lambda url: len(url) > 50 and url.count('/') > 5
+lambda url: any(domain in url for domain in ['api.', 'data.', 'feed.'])
+```
+
+#### Mixed Lists with AND/OR Logic
+```python
+# Combine multiple conditions
+CrawlerRunConfig(
+    url_matcher=[
+        "https://*",                        # Must be HTTPS
+        lambda url: 'internal' in url,      # Must contain 'internal'
+        lambda url: not url.endswith('.pdf') # Must not be PDF
+    ],
+    match_mode=MatchMode.AND  # ALL conditions must match
+)
+```
+
+### 6.3 Practical Example: News Site Crawler
+
+```python
+async def crawl_news_site():
+    dispatcher = MemoryAdaptiveDispatcher(
+        memory_threshold_percent=70.0,
+        rate_limiter=RateLimiter(base_delay=(1.0, 2.0))
+    )
+    
+    configs = [
+        # Homepage - light extraction
+        CrawlerRunConfig(
+            url_matcher=lambda url: url.rstrip('/') == 'https://news.ycombinator.com',
+            css_selector="nav, .headline",
+            extraction_strategy=None
+        ),
+        
+        # Article pages - full extraction
+        CrawlerRunConfig(
+            url_matcher="*/article/*",
+            extraction_strategy=CosineStrategy(
+                semantic_filter="article content",
+                word_count_threshold=100
+            ),
+            screenshot=True,
+            excluded_tags=["nav", "aside", "footer"]
+        ),
+        
+        # Author pages - metadata focus
+        CrawlerRunConfig(
+            url_matcher="*/author/*",
+            extraction_strategy=JsonCssExtractionStrategy({
+                "name": "h1.author-name",
+                "bio": ".author-bio",
+                "articles": "article.post-card h2"
+            })
+        ),
+        
+        # Everything else
+        CrawlerRunConfig()
+    ]
+    
+    async with AsyncWebCrawler() as crawler:
+        results = await crawler.arun_many(
+            urls=news_urls,
+            config=configs,
+            dispatcher=dispatcher
+        )
+```
+
+### 6.4 Best Practices
+
+1. **Order Matters**: Configs are evaluated in order - put specific patterns before general ones
+2. **Default Config Behavior**: 
+   - A config without `url_matcher` matches ALL URLs
+   - Always include a default config as the last item if you want to handle all URLs
+   - Without a default config, unmatched URLs will fail with "No matching configuration found"
+3. **Test Your Patterns**: Use the config's `is_match()` method to test patterns:
+   ```python
+   config = CrawlerRunConfig(url_matcher="*.pdf")
+   print(config.is_match("https://example.com/doc.pdf"))  # True
+   
+   default_config = CrawlerRunConfig()  # No url_matcher
+   print(default_config.is_match("https://any-url.com"))  # True - matches everything!
+   ```
+4. **Optimize for Performance**: 
+   - Disable JS for static content
+   - Skip screenshots for data APIs
+   - Use appropriate extraction strategies
+
+## 7. Summary
 
 1. **Two Dispatcher Types**:
 

docs/md_v2/advanced/pdf-parsing.md

@@ -0,0 +1,201 @@
+# PDF Processing Strategies
+
+Crawl4AI provides specialized strategies for handling and extracting content from PDF files. These strategies allow you to seamlessly integrate PDF processing into your crawling workflows, whether the PDFs are hosted online or stored locally.
+
+## `PDFCrawlerStrategy`
+
+### Overview
+`PDFCrawlerStrategy` is an implementation of `AsyncCrawlerStrategy` designed specifically for PDF documents. Instead of interpreting the input URL as an HTML webpage, this strategy treats it as a pointer to a PDF file. It doesn't perform deep crawling or HTML parsing itself but rather prepares the PDF source for a dedicated PDF scraping strategy. Its primary role is to identify the PDF source (web URL or local file) and pass it along the processing pipeline in a way that `AsyncWebCrawler` can handle.
+
+### When to Use
+Use `PDFCrawlerStrategy` when you need to:
+- Process PDF files using the `AsyncWebCrawler`.
+- Handle PDFs from both web URLs (e.g., `https://example.com/document.pdf`) and local file paths (e.g., `file:///path/to/your/document.pdf`).
+- Integrate PDF content extraction into a unified `CrawlResult` object, allowing consistent handling of PDF data alongside web page data.
+
+### Key Methods and Their Behavior
+-   **`__init__(self, logger: AsyncLogger = None)`**:
+    -   Initializes the strategy.
+    -   `logger`: An optional `AsyncLogger` instance (from `crawl4ai.async_logger`) for logging purposes.
+-   **`async crawl(self, url: str, **kwargs) -> AsyncCrawlResponse`**:
+    -   This method is called by the `AsyncWebCrawler` during the `arun` process.
+    -   It takes the `url` (which should point to a PDF) and creates a minimal `AsyncCrawlResponse`.
+    -   The `html` attribute of this response is typically empty or a placeholder, as the actual PDF content processing is deferred to the `PDFContentScrapingStrategy` (or a similar PDF-aware scraping strategy).
+    -   It sets `response_headers` to indicate "application/pdf" and `status_code` to 200.
+-   **`async close(self)`**:
+    -   A method for cleaning up any resources used by the strategy. For `PDFCrawlerStrategy`, this is usually minimal.
+-   **`async __aenter__(self)` / `async __aexit__(self, exc_type, exc_val, exc_tb)`**:
+    -   Enables asynchronous context management for the strategy, allowing it to be used with `async with`.
+
+### Example Usage
+```python
+import asyncio
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
+from crawl4ai.processors.pdf import PDFCrawlerStrategy, PDFContentScrapingStrategy
+
+async def main():
+    # Initialize the PDF crawler strategy
+    pdf_crawler_strategy = PDFCrawlerStrategy()
+
+    # PDFCrawlerStrategy is typically used in conjunction with PDFContentScrapingStrategy
+    # The scraping strategy handles the actual PDF content extraction
+    pdf_scraping_strategy = PDFContentScrapingStrategy()
+    run_config = CrawlerRunConfig(scraping_strategy=pdf_scraping_strategy)
+
+    async with AsyncWebCrawler(crawler_strategy=pdf_crawler_strategy) as crawler:
+        # Example with a remote PDF URL
+        pdf_url = "https://arxiv.org/pdf/2310.06825.pdf" # A public PDF from arXiv
+        
+        print(f"Attempting to process PDF: {pdf_url}")
+        result = await crawler.arun(url=pdf_url, config=run_config)
+
+        if result.success:
+            print(f"Successfully processed PDF: {result.url}")
+            print(f"Metadata Title: {result.metadata.get('title', 'N/A')}")
+            # Further processing of result.markdown, result.media, etc.
+            # would be done here, based on what PDFContentScrapingStrategy extracts.
+            if result.markdown and hasattr(result.markdown, 'raw_markdown'):
+                print(f"Extracted text (first 200 chars): {result.markdown.raw_markdown[:200]}...")
+            else:
+                print("No markdown (text) content extracted.")
+        else:
+            print(f"Failed to process PDF: {result.error_message}")
+
+if __name__ == "__main__":
+    asyncio.run(main())
+```
+
+### Pros and Cons
+**Pros:**
+-   Enables `AsyncWebCrawler` to handle PDF sources directly using familiar `arun` calls.
+-   Provides a consistent interface for specifying PDF sources (URLs or local paths).
+-   Abstracts the source handling, allowing a separate scraping strategy to focus on PDF content parsing.
+
+**Cons:**
+-   Does not perform any PDF data extraction itself; it strictly relies on a compatible scraping strategy (like `PDFContentScrapingStrategy`) to process the PDF.
+-   Has limited utility on its own; most of its value comes from being paired with a PDF-specific content scraping strategy.
+
+---
+
+## `PDFContentScrapingStrategy`
+
+### Overview
+`PDFContentScrapingStrategy` is an implementation of `ContentScrapingStrategy` designed to extract text, metadata, and optionally images from PDF documents. It is intended to be used in conjunction with a crawler strategy that can provide it with a PDF source, such as `PDFCrawlerStrategy`. This strategy uses the `NaivePDFProcessorStrategy` internally to perform the low-level PDF parsing.
+
+### When to Use
+Use `PDFContentScrapingStrategy` when your `AsyncWebCrawler` (often configured with `PDFCrawlerStrategy`) needs to:
+-   Extract textual content page by page from a PDF document.
+-   Retrieve standard metadata embedded within the PDF (e.g., title, author, subject, creation date, page count).
+-   Optionally, extract images contained within the PDF pages. These images can be saved to a local directory or made available for further processing.
+-   Produce a `ScrapingResult` that can be converted into a `CrawlResult`, making PDF content accessible in a manner similar to HTML web content (e.g., text in `result.markdown`, metadata in `result.metadata`).
+
+### Key Configuration Attributes
+When initializing `PDFContentScrapingStrategy`, you can configure its behavior using the following attributes:
+-   **`extract_images: bool = False`**: If `True`, the strategy will attempt to extract images from the PDF.
+-   **`save_images_locally: bool = False`**: If `True` (and `extract_images` is also `True`), extracted images will be saved to disk in the `image_save_dir`. If `False`, image data might be available in another form (e.g., base64, depending on the underlying processor) but not saved as separate files by this strategy.
+-   **`image_save_dir: str = None`**: Specifies the directory where extracted images should be saved if `save_images_locally` is `True`. If `None`, a default or temporary directory might be used.
+-   **`batch_size: int = 4`**: Defines how many PDF pages are processed in a single batch. This can be useful for managing memory when dealing with very large PDF documents.
+-   **`logger: AsyncLogger = None`**: An optional `AsyncLogger` instance for logging.
+
+### Key Methods and Their Behavior
+-   **`__init__(self, save_images_locally: bool = False, extract_images: bool = False, image_save_dir: str = None, batch_size: int = 4, logger: AsyncLogger = None)`**:
+    -   Initializes the strategy with configurations for image handling, batch processing, and logging. It sets up an internal `NaivePDFProcessorStrategy` instance which performs the actual PDF parsing.
+-   **`scrap(self, url: str, html: str, **params) -> ScrapingResult`**:
+    -   This is the primary synchronous method called by the crawler (via `ascrap`) to process the PDF.
+    -   `url`: The path or URL to the PDF file (provided by `PDFCrawlerStrategy` or similar).
+    -   `html`: Typically an empty string when used with `PDFCrawlerStrategy`, as the content is a PDF, not HTML.
+    -   It first ensures the PDF is accessible locally (downloads it to a temporary file if `url` is remote).
+    -   It then uses its internal PDF processor to extract text, metadata, and images (if configured).
+    -   The extracted information is compiled into a `ScrapingResult` object:
+        -   `cleaned_html`: Contains an HTML-like representation of the PDF, where each page's content is often wrapped in a `<div>` with page number information.
+        -   `media`: A dictionary where `media["images"]` will contain information about extracted images if `extract_images` was `True`.
+        -   `links`: A dictionary where `links["urls"]` can contain URLs found within the PDF content.
+        -   `metadata`: A dictionary holding PDF metadata (e.g., title, author, num_pages).
+-   **`async ascrap(self, url: str, html: str, **kwargs) -> ScrapingResult`**:
+    -   The asynchronous version of `scrap`. Under the hood, it typically runs the synchronous `scrap` method in a separate thread using `asyncio.to_thread` to avoid blocking the event loop.
+-   **`_get_pdf_path(self, url: str) -> str`**:
+    -   A private helper method to manage PDF file access. If the `url` is remote (http/https), it downloads the PDF to a temporary local file and returns its path. If `url` indicates a local file (`file://` or a direct path), it resolves and returns the local path.
+
+### Example Usage
+```python
+import asyncio
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
+from crawl4ai.processors.pdf import PDFCrawlerStrategy, PDFContentScrapingStrategy
+import os # For creating image directory
+
+async def main():
+    # Define the directory for saving extracted images
+    image_output_dir = "./my_pdf_images"
+    os.makedirs(image_output_dir, exist_ok=True)
+
+    # Configure the PDF content scraping strategy
+    # Enable image extraction and specify where to save them
+    pdf_scraping_cfg = PDFContentScrapingStrategy(
+        extract_images=True,
+        save_images_locally=True,
+        image_save_dir=image_output_dir,
+        batch_size=2 # Process 2 pages at a time for demonstration
+    )
+
+    # The PDFCrawlerStrategy is needed to tell AsyncWebCrawler how to "crawl" a PDF
+    pdf_crawler_cfg = PDFCrawlerStrategy()
+
+    # Configure the overall crawl run
+    run_cfg = CrawlerRunConfig(
+        scraping_strategy=pdf_scraping_cfg # Use our PDF scraping strategy
+    )
+
+    # Initialize the crawler with the PDF-specific crawler strategy
+    async with AsyncWebCrawler(crawler_strategy=pdf_crawler_cfg) as crawler:
+        pdf_url = "https://arxiv.org/pdf/2310.06825.pdf" # Example PDF
+        
+        print(f"Starting PDF processing for: {pdf_url}")
+        result = await crawler.arun(url=pdf_url, config=run_cfg)
+
+        if result.success:
+            print("\n--- PDF Processing Successful ---")
+            print(f"Processed URL: {result.url}")
+            
+            print("\n--- Metadata ---")
+            for key, value in result.metadata.items():
+                print(f"  {key.replace('_', ' ').title()}: {value}")
+
+            if result.markdown and hasattr(result.markdown, 'raw_markdown'):
+                print(f"\n--- Extracted Text (Markdown Snippet) ---")
+                print(result.markdown.raw_markdown[:500].strip() + "...")
+            else:
+                print("\nNo text (markdown) content extracted.")
+
+            if result.media and result.media.get("images"):
+                print(f"\n--- Image Extraction ---")
+                print(f"Extracted {len(result.media['images'])} image(s).")
+                for i, img_info in enumerate(result.media["images"][:2]): # Show info for first 2 images
+                    print(f"  Image {i+1}:")
+                    print(f"    Page: {img_info.get('page')}")
+                    print(f"    Format: {img_info.get('format', 'N/A')}")
+                    if img_info.get('path'):
+                        print(f"    Saved at: {img_info.get('path')}")
+            else:
+                print("\nNo images were extracted (or extract_images was False).")
+        else:
+            print(f"\n--- PDF Processing Failed ---")
+            print(f"Error: {result.error_message}")
+
+if __name__ == "__main__":
+    asyncio.run(main())
+```
+
+### Pros and Cons
+
+**Pros:**
+-   Provides a comprehensive way to extract text, metadata, and (optionally) images from PDF documents.
+-   Handles both remote PDFs (via URL) and local PDF files.
+-   Configurable image extraction allows saving images to disk or accessing their data.
+-   Integrates smoothly with the `CrawlResult` object structure, making PDF-derived data accessible in a way consistent with web-scraped data.
+-   The `batch_size` parameter can help in managing memory consumption when processing large or numerous PDF pages.
+
+**Cons:**
+-   Extraction quality and performance can vary significantly depending on the PDF's complexity, encoding, and whether it's image-based (scanned) or text-based.
+-   Image extraction can be resource-intensive (both CPU and disk space if `save_images_locally` is true).
+-   Relies on `NaivePDFProcessorStrategy` internally, which might have limitations with very complex layouts, encrypted PDFs, or forms compared to more sophisticated PDF parsing libraries. Scanned PDFs will not yield text unless an OCR step is performed (which is not part of this strategy by default).
+-   Link extraction from PDFs can be basic and depends on how hyperlinks are embedded in the document.

docs/md_v2/advanced/proxy-security.md

@@ -25,44 +25,70 @@ Use an authenticated proxy with `BrowserConfig`:
 ```python
 from crawl4ai.async_configs import BrowserConfig
 
-proxy_config = {
-    "server": "http://proxy.example.com:8080",
-    "username": "user",
-    "password": "pass"
-}
-
-browser_config = BrowserConfig(proxy_config=proxy_config)
+browser_config = BrowserConfig(proxy="http://[username]:[password]@[host]:[port]")
 async with AsyncWebCrawler(config=browser_config) as crawler:
     result = await crawler.arun(url="https://example.com")
 ```
 
-Here's the corrected documentation:
 
 ## Rotating Proxies 
 
 Example using a proxy rotation service dynamically:
 
 ```python
-from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
-
-async def get_next_proxy():
-    # Your proxy rotation logic here
-    return {"server": "http://next.proxy.com:8080"}
-
+import re
+from crawl4ai import (
+    AsyncWebCrawler,
+    BrowserConfig,
+    CrawlerRunConfig,
+    CacheMode,
+    RoundRobinProxyStrategy,
+)
+import asyncio
+from crawl4ai import ProxyConfig
 async def main():
-    browser_config = BrowserConfig()
-    run_config = CrawlerRunConfig()
-    
-    async with AsyncWebCrawler(config=browser_config) as crawler:
-        # For each URL, create a new run config with different proxy
-        for url in urls:
-            proxy = await get_next_proxy()
-            # Clone the config and update proxy - this creates a new browser context
-            current_config = run_config.clone(proxy_config=proxy)
-            result = await crawler.arun(url=url, config=current_config)
+    # Load proxies and create rotation strategy
+    proxies = ProxyConfig.from_env()
+    #eg: export PROXIES="ip1:port1:username1:password1,ip2:port2:username2:password2"
+    if not proxies:
+        print("No proxies found in environment. Set PROXIES env variable!")
+        return
+
+    proxy_strategy = RoundRobinProxyStrategy(proxies)
+
+    # Create configs
+    browser_config = BrowserConfig(headless=True, verbose=False)
+    run_config = CrawlerRunConfig(
+        cache_mode=CacheMode.BYPASS,
+        proxy_rotation_strategy=proxy_strategy
+    )
+
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        urls = ["https://httpbin.org/ip"] * (len(proxies) * 2)  # Test each proxy twice
+
+        print("\n📈 Initializing crawler with proxy rotation...")
+        async with AsyncWebCrawler(config=browser_config) as crawler:
+            print("\n🚀 Starting batch crawl with proxy rotation...")
+            results = await crawler.arun_many(
+                urls=urls,
+                config=run_config
+            )
+            for result in results:
+                if result.success:
+                    ip_match = re.search(r'(?:[0-9]{1,3}\.){3}[0-9]{1,3}', result.html)
+                    current_proxy = run_config.proxy_config if run_config.proxy_config else None
+
+                    if current_proxy and ip_match:
+                        print(f"URL {result.url}")
+                        print(f"Proxy {current_proxy.server} -> Response IP: {ip_match.group(0)}")
+                        verified = ip_match.group(0) == current_proxy.ip
+                        if verified:
+                            print(f"✅ Proxy working! IP matches: {current_proxy.ip}")
+                        else:
+                            print("❌ Proxy failed or IP mismatch!")
+                    print("---")
+
+asyncio.run(main())
 
-if __name__ == "__main__":
-    import asyncio
-    asyncio.run(main())
 ```
 

docs/md_v2/advanced/session-management.md

@@ -49,46 +49,75 @@ from crawl4ai import JsonCssExtractionStrategy
 from crawl4ai.cache_context import CacheMode
 
 async def crawl_dynamic_content():
-    async with AsyncWebCrawler() as crawler:
-        session_id = "github_commits_session"
-        url = "https://github.com/microsoft/TypeScript/commits/main"
-        all_commits = []
+    url = "https://github.com/microsoft/TypeScript/commits/main"
+    session_id = "wait_for_session"
+    all_commits = []
 
-        # Define extraction schema
-        schema = {
-            "name": "Commit Extractor",
-            "baseSelector": "li.Box-sc-g0xbh4-0",
-            "fields": [{
-                "name": "title", "selector": "h4.markdown-title", "type": "text"
-            }],
-        }
-        extraction_strategy = JsonCssExtractionStrategy(schema)
+    js_next_page = """
+    const commits = document.querySelectorAll('li[data-testid="commit-row-item"] h4');
+    if (commits.length > 0) {
+        window.lastCommit = commits[0].textContent.trim();
+    }
+    const button = document.querySelector('a[data-testid="pagination-next-button"]');
+    if (button) {button.click(); console.log('button clicked') }
+    """
 
-        # JavaScript and wait configurations
-        js_next_page = """document.querySelector('a[data-testid="pagination-next-button"]').click();"""
-        wait_for = """() => document.querySelectorAll('li.Box-sc-g0xbh4-0').length > 0"""
-
-        # Crawl multiple pages
+    wait_for = """() => {
+        const commits = document.querySelectorAll('li[data-testid="commit-row-item"] h4');
+        if (commits.length === 0) return false;
+        const firstCommit = commits[0].textContent.trim();
+        return firstCommit !== window.lastCommit;
+    }"""
+    
+    schema = {
+        "name": "Commit Extractor",
+        "baseSelector": "li[data-testid='commit-row-item']",
+        "fields": [
+            {
+                "name": "title",
+                "selector": "h4 a",
+                "type": "text",
+                "transform": "strip",
+            },
+        ],
+    }
+    extraction_strategy = JsonCssExtractionStrategy(schema, verbose=True)
+    
+    
+    browser_config = BrowserConfig(
+        verbose=True,
+        headless=False,
+    )
+        
+    async with AsyncWebCrawler(config=browser_config) as crawler:
         for page in range(3):
-            config = CrawlerRunConfig(
-                url=url,
+            crawler_config = CrawlerRunConfig(
                 session_id=session_id,
+                css_selector="li[data-testid='commit-row-item']",
                 extraction_strategy=extraction_strategy,
                 js_code=js_next_page if page > 0 else None,
                 wait_for=wait_for if page > 0 else None,
                 js_only=page > 0,
-                cache_mode=CacheMode.BYPASS
+                cache_mode=CacheMode.BYPASS,
+                capture_console_messages=True,
             )
-
-            result = await crawler.arun(config=config)
-            if result.success:
+            
+            result = await crawler.arun(url=url, config=crawler_config)
+            
+            if result.console_messages:
+                print(f"Page {page + 1} console messages:", result.console_messages)
+            
+            if result.extracted_content:
+                # print(f"Page {page + 1} result:", result.extracted_content)
                 commits = json.loads(result.extracted_content)
                 all_commits.extend(commits)
                 print(f"Page {page + 1}: Found {len(commits)} commits")
+            else:
+                print(f"Page {page + 1}: No content extracted")
 
+        print(f"Successfully crawled {len(all_commits)} commits across 3 pages")
         # Clean up session
         await crawler.crawler_strategy.kill_session(session_id)
-        return all_commits
 ```
 
 ---

docs/md_v2/advanced/undetected-browser.md

@@ -0,0 +1,394 @@
+# Undetected Browser Mode
+
+## Overview
+
+Crawl4AI offers two powerful anti-bot features to help you access websites with bot detection:
+
+1. **Stealth Mode** - Uses playwright-stealth to modify browser fingerprints and behaviors
+2. **Undetected Browser Mode** - Advanced browser adapter with deep-level patches for sophisticated bot detection
+
+This guide covers both features and helps you choose the right approach for your needs.
+
+## Anti-Bot Features Comparison
+
+| Feature | Regular Browser | Stealth Mode | Undetected Browser |
+|---------|----------------|--------------|-------------------|
+| WebDriver Detection | ❌ | ✅ | ✅ |
+| Navigator Properties | ❌ | ✅ | ✅ |
+| Plugin Emulation | ❌ | ✅ | ✅ |
+| CDP Detection | ❌ | Partial | ✅ |
+| Deep Browser Patches | ❌ | ❌ | ✅ |
+| Performance Impact | None | Minimal | Moderate |
+| Setup Complexity | None | None | Minimal |
+
+## When to Use Each Approach
+
+### Use Regular Browser + Stealth Mode When:
+- Sites have basic bot detection (checking navigator.webdriver, plugins, etc.)
+- You need good performance with basic protection
+- Sites check for common automation indicators
+
+### Use Undetected Browser When:
+- Sites employ sophisticated bot detection services (Cloudflare, DataDome, etc.)
+- Stealth mode alone isn't sufficient
+- You're willing to trade some performance for better evasion
+
+### Best Practice: Progressive Enhancement
+1. **Start with**: Regular browser + Stealth mode
+2. **If blocked**: Switch to Undetected browser
+3. **If still blocked**: Combine Undetected browser + Stealth mode
+
+## Stealth Mode
+
+Stealth mode is the simpler anti-bot solution that works with both regular and undetected browsers:
+
+```python
+from crawl4ai import AsyncWebCrawler, BrowserConfig
+
+# Enable stealth mode with regular browser
+browser_config = BrowserConfig(
+    enable_stealth=True,  # Simple flag to enable
+    headless=False       # Better for avoiding detection
+)
+
+async with AsyncWebCrawler(config=browser_config) as crawler:
+    result = await crawler.arun("https://example.com")
+```
+
+### What Stealth Mode Does:
+- Removes `navigator.webdriver` flag
+- Modifies browser fingerprints
+- Emulates realistic plugin behavior
+- Adjusts navigator properties
+- Fixes common automation leaks
+
+## Undetected Browser Mode
+
+For sites with sophisticated bot detection that stealth mode can't bypass, use the undetected browser adapter:
+
+### Key Features
+
+- **Drop-in Replacement**: Uses the same API as regular browser mode
+- **Enhanced Stealth**: Built-in patches to evade common detection methods
+- **Browser Adapter Pattern**: Seamlessly switch between regular and undetected modes
+- **Automatic Installation**: `crawl4ai-setup` installs all necessary browser dependencies
+
+### Quick Start
+
+```python
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler, 
+    BrowserConfig, 
+    CrawlerRunConfig,
+    UndetectedAdapter
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+async def main():
+    # Create the undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    
+    # Create browser config
+    browser_config = BrowserConfig(
+        headless=False,  # Headless mode can be detected easier
+        verbose=True,
+    )
+    
+    # Create the crawler strategy with undetected adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    # Create the crawler with our custom strategy
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        # Your crawling code here
+        result = await crawler.arun(
+            url="https://example.com",
+            config=CrawlerRunConfig()
+        )
+        print(result.markdown[:500])
+
+asyncio.run(main())
+```
+
+## Combining Both Features
+
+For maximum evasion, combine stealth mode with undetected browser:
+
+```python
+from crawl4ai import AsyncWebCrawler, BrowserConfig, UndetectedAdapter
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+# Create browser config with stealth enabled
+browser_config = BrowserConfig(
+    enable_stealth=True,  # Enable stealth mode
+    headless=False
+)
+
+# Create undetected adapter
+adapter = UndetectedAdapter()
+
+# Create strategy with both features
+strategy = AsyncPlaywrightCrawlerStrategy(
+    browser_config=browser_config,
+    browser_adapter=adapter
+)
+
+async with AsyncWebCrawler(
+    crawler_strategy=strategy,
+    config=browser_config
+) as crawler:
+    result = await crawler.arun("https://protected-site.com")
+```
+
+## Examples
+
+### Example 1: Basic Stealth Mode
+
+```python
+import asyncio
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
+
+async def test_stealth_mode():
+    # Simple stealth mode configuration
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun(
+            url="https://bot.sannysoft.com",
+            config=CrawlerRunConfig(screenshot=True)
+        )
+        
+        if result.success:
+            print("✓ Successfully accessed bot detection test site")
+            # Save screenshot to verify detection results
+            if result.screenshot:
+                import base64
+                with open("stealth_test.png", "wb") as f:
+                    f.write(base64.b64decode(result.screenshot))
+                print("✓ Screenshot saved - check for green (passed) tests")
+
+asyncio.run(test_stealth_mode())
+```
+
+### Example 2: Undetected Browser Mode
+
+```python
+import asyncio
+from crawl4ai import (
+    AsyncWebCrawler,
+    BrowserConfig,
+    CrawlerRunConfig,
+    UndetectedAdapter
+)
+from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+
+
+async def main():
+    # Create browser config
+    browser_config = BrowserConfig(
+        headless=False,
+        verbose=True,
+    )
+    
+    # Create the undetected adapter
+    undetected_adapter = UndetectedAdapter()
+    
+    # Create the crawler strategy with the undetected adapter
+    crawler_strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=undetected_adapter
+    )
+    
+    # Create the crawler with our custom strategy
+    async with AsyncWebCrawler(
+        crawler_strategy=crawler_strategy,
+        config=browser_config
+    ) as crawler:
+        # Configure the crawl
+        crawler_config = CrawlerRunConfig(
+            markdown_generator=DefaultMarkdownGenerator(
+                content_filter=PruningContentFilter()
+            ),
+            capture_console_messages=True,  # Test adapter console capture
+        )
+        
+        # Test on a site that typically detects bots
+        print("Testing undetected adapter...")
+        result: CrawlResult = await crawler.arun(
+            url="https://www.helloworld.org", 
+            config=crawler_config
+        )
+        
+        print(f"Status: {result.status_code}")
+        print(f"Success: {result.success}")
+        print(f"Console messages captured: {len(result.console_messages or [])}")
+        print(f"Markdown content (first 500 chars):\n{result.markdown.raw_markdown[:500]}")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())
+```
+
+## Browser Adapter Pattern
+
+The undetected browser support is implemented using an adapter pattern, allowing seamless switching between different browser implementations:
+
+```python
+# Regular browser adapter (default)
+from crawl4ai import PlaywrightAdapter
+regular_adapter = PlaywrightAdapter()
+
+# Undetected browser adapter
+from crawl4ai import UndetectedAdapter
+undetected_adapter = UndetectedAdapter()
+```
+
+The adapter handles:
+- JavaScript execution
+- Console message capture
+- Error handling
+- Browser-specific optimizations
+
+## Best Practices
+
+1. **Avoid Headless Mode**: Detection is easier in headless mode
+   ```python
+   browser_config = BrowserConfig(headless=False)
+   ```
+
+2. **Use Reasonable Delays**: Don't rush through pages
+   ```python
+   crawler_config = CrawlerRunConfig(
+       wait_time=3.0,  # Wait 3 seconds after page load
+       delay_before_return_html=2.0  # Additional delay
+   )
+   ```
+
+3. **Rotate User Agents**: You can customize user agents
+   ```python
+   browser_config = BrowserConfig(
+       headers={"User-Agent": "your-user-agent"}
+   )
+   ```
+
+4. **Handle Failures Gracefully**: Some sites may still detect and block
+   ```python
+   if not result.success:
+       print(f"Crawl failed: {result.error_message}")
+   ```
+
+## Advanced Usage Tips
+
+### Progressive Detection Handling
+
+```python
+async def crawl_with_progressive_evasion(url):
+    # Step 1: Try regular browser with stealth
+    browser_config = BrowserConfig(
+        enable_stealth=True,
+        headless=False
+    )
+    
+    async with AsyncWebCrawler(config=browser_config) as crawler:
+        result = await crawler.arun(url)
+        if result.success and "Access Denied" not in result.html:
+            return result
+    
+    # Step 2: If blocked, try undetected browser
+    print("Regular + stealth blocked, trying undetected browser...")
+    
+    adapter = UndetectedAdapter()
+    strategy = AsyncPlaywrightCrawlerStrategy(
+        browser_config=browser_config,
+        browser_adapter=adapter
+    )
+    
+    async with AsyncWebCrawler(
+        crawler_strategy=strategy,
+        config=browser_config
+    ) as crawler:
+        result = await crawler.arun(url)
+        return result
+```
+
+## Installation
+
+The undetected browser dependencies are automatically installed when you run:
+
+```bash
+crawl4ai-setup
+```
+
+This command installs all necessary browser dependencies for both regular and undetected modes.
+
+## Limitations
+
+- **Performance**: Slightly slower than regular mode due to additional patches
+- **Headless Detection**: Some sites can still detect headless mode
+- **Resource Usage**: May use more resources than regular mode
+- **Not 100% Guaranteed**: Advanced anti-bot services are constantly evolving
+
+## Troubleshooting
+
+### Browser Not Found
+
+Run the setup command:
+```bash
+crawl4ai-setup
+```
+
+### Detection Still Occurring
+
+Try combining with other features:
+```python
+crawler_config = CrawlerRunConfig(
+    simulate_user=True,  # Add user simulation
+    magic=True,  # Enable magic mode
+    wait_time=5.0,  # Longer waits
+)
+```
+
+### Performance Issues
+
+If experiencing slow performance:
+```python
+# Use selective undetected mode only for protected sites
+if is_protected_site(url):
+    adapter = UndetectedAdapter()
+else:
+    adapter = PlaywrightAdapter()  # Default adapter
+```
+
+## Future Plans
+
+**Note**: In future versions of Crawl4AI, we may enable stealth mode and undetected browser by default to provide better out-of-the-box success rates. For now, users should explicitly enable these features when needed.
+
+## Conclusion
+
+Crawl4AI provides flexible anti-bot solutions:
+
+1. **Start Simple**: Use regular browser + stealth mode for most sites
+2. **Escalate if Needed**: Switch to undetected browser for sophisticated protection
+3. **Combine for Maximum Effect**: Use both features together when facing the toughest challenges
+
+Remember:
+- Always respect robots.txt and website terms of service
+- Use appropriate delays to avoid overwhelming servers
+- Consider the performance trade-offs of each approach
+- Test progressively to find the minimum necessary evasion level
+
+## See Also
+
+- [Advanced Features](advanced-features.md) - Overview of all advanced features
+- [Proxy & Security](proxy-security.md) - Using proxies with anti-bot features
+- [Session Management](session-management.md) - Maintaining sessions across requests
+- [Identity Based Crawling](identity-based-crawling.md) - Additional anti-detection strategies

docs/md_v2/advanced/virtual-scroll.md

@@ -91,13 +91,12 @@ async def crawl_twitter_timeline():
         wait_after_scroll=1.0  # Twitter needs time to load
     )
     
+    browser_config = BrowserConfig(headless=True)  # Set to False to watch it work
     config = CrawlerRunConfig(
-        virtual_scroll_config=virtual_config,
-        # Optional: Set headless=False to watch it work
-        # browser_config=BrowserConfig(headless=False)
+        virtual_scroll_config=virtual_config
     )
     
-    async with AsyncWebCrawler() as crawler:
+    async with AsyncWebCrawler(config=browser_config) as crawler:
         result = await crawler.arun(
             url="https://twitter.com/search?q=AI",
             config=config
@@ -200,7 +199,7 @@ Use **scan_full_page** when:
 Virtual Scroll works seamlessly with extraction strategies:
 
 ```python
-from crawl4ai import LLMExtractionStrategy
+from crawl4ai import LLMExtractionStrategy, LLMConfig
 
 # Define extraction schema
 schema = {
@@ -222,7 +221,7 @@ config = CrawlerRunConfig(
         scroll_count=20
     ),
     extraction_strategy=LLMExtractionStrategy(
-        provider="openai/gpt-4o-mini",
+        llm_config=LLMConfig(provider="openai/gpt-4o-mini"),
         schema=schema
     )
 )

docs/md_v2/api/adaptive-crawler.md

@@ -0,0 +1,244 @@
+# AdaptiveCrawler
+
+The `AdaptiveCrawler` class implements intelligent web crawling that automatically determines when sufficient information has been gathered to answer a query. It uses a three-layer scoring system to evaluate coverage, consistency, and saturation.
+
+## Constructor
+
+```python
+AdaptiveCrawler(
+    crawler: AsyncWebCrawler,
+    config: Optional[AdaptiveConfig] = None
+)
+```
+
+### Parameters
+
+- **crawler** (`AsyncWebCrawler`): The underlying web crawler instance to use for fetching pages
+- **config** (`Optional[AdaptiveConfig]`): Configuration settings for adaptive crawling behavior. If not provided, uses default settings.
+
+## Primary Method
+
+### digest()
+
+The main method that performs adaptive crawling starting from a URL with a specific query.
+
+```python
+async def digest(
+    start_url: str,
+    query: str,
+    resume_from: Optional[Union[str, Path]] = None
+) -> CrawlState
+```
+
+#### Parameters
+
+- **start_url** (`str`): The starting URL for crawling
+- **query** (`str`): The search query that guides the crawling process
+- **resume_from** (`Optional[Union[str, Path]]`): Path to a saved state file to resume from
+
+#### Returns
+
+- **CrawlState**: The final crawl state containing all crawled URLs, knowledge base, and metrics
+
+#### Example
+
+```python
+async with AsyncWebCrawler() as crawler:
+    adaptive = AdaptiveCrawler(crawler)
+    state = await adaptive.digest(
+        start_url="https://docs.python.org",
+        query="async context managers"
+    )
+```
+
+## Properties
+
+### confidence
+
+Current confidence score (0-1) indicating information sufficiency.
+
+```python
+@property
+def confidence(self) -> float
+```
+
+### coverage_stats
+
+Dictionary containing detailed coverage statistics.
+
+```python
+@property  
+def coverage_stats(self) -> Dict[str, float]
+```
+
+Returns:
+- **coverage**: Query term coverage score
+- **consistency**: Information consistency score  
+- **saturation**: Content saturation score
+- **confidence**: Overall confidence score
+
+### is_sufficient
+
+Boolean indicating whether sufficient information has been gathered.
+
+```python
+@property
+def is_sufficient(self) -> bool
+```
+
+### state
+
+Access to the current crawl state.
+
+```python
+@property
+def state(self) -> CrawlState
+```
+
+## Methods
+
+### get_relevant_content()
+
+Retrieve the most relevant content from the knowledge base.
+
+```python
+def get_relevant_content(
+    self,
+    top_k: int = 5
+) -> List[Dict[str, Any]]
+```
+
+#### Parameters
+
+- **top_k** (`int`): Number of top relevant documents to return (default: 5)
+
+#### Returns
+
+List of dictionaries containing:
+- **url**: The URL of the page
+- **content**: The page content
+- **score**: Relevance score
+- **metadata**: Additional page metadata
+
+### print_stats()
+
+Display crawl statistics in formatted output.
+
+```python
+def print_stats(
+    self,
+    detailed: bool = False
+) -> None
+```
+
+#### Parameters
+
+- **detailed** (`bool`): If True, shows detailed metrics with colors. If False, shows summary table.
+
+### export_knowledge_base()
+
+Export the collected knowledge base to a JSONL file.
+
+```python
+def export_knowledge_base(
+    self,
+    path: Union[str, Path]
+) -> None
+```
+
+#### Parameters
+
+- **path** (`Union[str, Path]`): Output file path for JSONL export
+
+#### Example
+
+```python
+adaptive.export_knowledge_base("my_knowledge.jsonl")
+```
+
+### import_knowledge_base()
+
+Import a previously exported knowledge base.
+
+```python
+def import_knowledge_base(
+    self,
+    path: Union[str, Path]
+) -> None
+```
+
+#### Parameters
+
+- **path** (`Union[str, Path]`): Path to JSONL file to import
+
+## Configuration
+
+The `AdaptiveConfig` class controls the behavior of adaptive crawling:
+
+```python
+@dataclass
+class AdaptiveConfig:
+    confidence_threshold: float = 0.8      # Stop when confidence reaches this
+    max_pages: int = 50                    # Maximum pages to crawl
+    top_k_links: int = 5                   # Links to follow per page
+    min_gain_threshold: float = 0.1        # Minimum expected gain to continue
+    save_state: bool = False               # Auto-save crawl state
+    state_path: Optional[str] = None       # Path for state persistence
+```
+
+### Example with Custom Config
+
+```python
+config = AdaptiveConfig(
+    confidence_threshold=0.7,
+    max_pages=20,
+    top_k_links=3
+)
+
+adaptive = AdaptiveCrawler(crawler, config=config)
+```
+
+## Complete Example
+
+```python
+import asyncio
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+async def main():
+    # Configure adaptive crawling
+    config = AdaptiveConfig(
+        confidence_threshold=0.75,
+        max_pages=15,
+        save_state=True,
+        state_path="my_crawl.json"
+    )
+    
+    async with AsyncWebCrawler() as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        # Start crawling
+        state = await adaptive.digest(
+            start_url="https://example.com/docs",
+            query="authentication oauth2 jwt"
+        )
+        
+        # Check results
+        print(f"Confidence achieved: {adaptive.confidence:.0%}")
+        adaptive.print_stats()
+        
+        # Get most relevant pages
+        for page in adaptive.get_relevant_content(top_k=3):
+            print(f"- {page['url']} (score: {page['score']:.2f})")
+        
+        # Export for later use
+        adaptive.export_knowledge_base("auth_knowledge.jsonl")
+
+if __name__ == "__main__":
+    asyncio.run(main())
+```
+
+## See Also
+
+- [digest() Method Reference](digest.md)
+- [Adaptive Crawling Guide](../core/adaptive-crawling.md)
+- [Advanced Adaptive Strategies](../advanced/adaptive-strategies.md)

docs/md_v2/api/arun_many.md

@@ -7,7 +7,7 @@
 ```python
 async def arun_many(
     urls: Union[List[str], List[Any]],
-    config: Optional[CrawlerRunConfig] = None,
+    config: Optional[Union[CrawlerRunConfig, List[CrawlerRunConfig]]] = None,
     dispatcher: Optional[BaseDispatcher] = None,
     ...
 ) -> Union[List[CrawlResult], AsyncGenerator[CrawlResult, None]]:
@@ -15,7 +15,9 @@ async def arun_many(
     Crawl multiple URLs concurrently or in batches.
 
     :param urls: A list of URLs (or tasks) to crawl.
-    :param config: (Optional) A default `CrawlerRunConfig` applying to each crawl.
+    :param config: (Optional) Either:
+        - A single `CrawlerRunConfig` applying to all URLs
+        - A list of `CrawlerRunConfig` objects with url_matcher patterns
     :param dispatcher: (Optional) A concurrency controller (e.g. MemoryAdaptiveDispatcher).
     ...
     :return: Either a list of `CrawlResult` objects, or an async generator if streaming is enabled.
@@ -95,10 +97,70 @@ results = await crawler.arun_many(
 )
 ```
 
+### URL-Specific Configurations
+
+Instead of using one config for all URLs, provide a list of configs with `url_matcher` patterns:
+
+```python
+from crawl4ai import CrawlerRunConfig, MatchMode
+from crawl4ai.processors.pdf import PDFContentScrapingStrategy
+from crawl4ai.extraction_strategy import JsonCssExtractionStrategy
+from crawl4ai.content_filter_strategy import PruningContentFilter
+from crawl4ai.markdown_generation_strategy import DefaultMarkdownGenerator
+
+# PDF files - specialized extraction
+pdf_config = CrawlerRunConfig(
+    url_matcher="*.pdf",
+    scraping_strategy=PDFContentScrapingStrategy()
+)
+
+# Blog/article pages - content filtering
+blog_config = CrawlerRunConfig(
+    url_matcher=["*/blog/*", "*/article/*", "*python.org*"],
+    markdown_generator=DefaultMarkdownGenerator(
+        content_filter=PruningContentFilter(threshold=0.48)
+    )
+)
+
+# Dynamic pages - JavaScript execution
+github_config = CrawlerRunConfig(
+    url_matcher=lambda url: 'github.com' in url,
+    js_code="window.scrollTo(0, 500);"
+)
+
+# API endpoints - JSON extraction
+api_config = CrawlerRunConfig(
+    url_matcher=lambda url: 'api' in url or url.endswith('.json'),
+    # Custome settings for JSON extraction
+)
+
+# Default fallback config
+default_config = CrawlerRunConfig()  # No url_matcher means it never matches except as fallback
+
+# Pass the list of configs - first match wins!
+results = await crawler.arun_many(
+    urls=[
+        "https://www.w3.org/WAI/ER/tests/xhtml/testfiles/resources/pdf/dummy.pdf",  # → pdf_config
+        "https://blog.python.org/",  # → blog_config
+        "https://github.com/microsoft/playwright",  # → github_config
+        "https://httpbin.org/json",  # → api_config
+        "https://example.com/"  # → default_config
+    ],
+    config=[pdf_config, blog_config, github_config, api_config, default_config]
+)
+```
+
+**URL Matching Features**:
+- **String patterns**: `"*.pdf"`, `"*/blog/*"`, `"*python.org*"`
+- **Function matchers**: `lambda url: 'api' in url`
+- **Mixed patterns**: Combine strings and functions with `MatchMode.OR` or `MatchMode.AND`
+- **First match wins**: Configs are evaluated in order
+
 **Key Points**:
 - Each URL is processed by the same or separate sessions, depending on the dispatcher’s strategy.
 - `dispatch_result` in each `CrawlResult` (if using concurrency) can hold memory and timing info.  
 - If you need to handle authentication or session IDs, pass them in each individual task or within your run config.
+- **Important**: Always include a default config (without `url_matcher`) as the last item if you want to handle all URLs. Otherwise, unmatched URLs will fail.
 
 ### Return Value
 

docs/md_v2/api/digest.md

@@ -0,0 +1,181 @@
+# digest()
+
+The `digest()` method is the primary interface for adaptive web crawling. It intelligently crawls websites starting from a given URL, guided by a query, and automatically determines when sufficient information has been gathered.
+
+## Method Signature
+
+```python
+async def digest(
+    start_url: str,
+    query: str,
+    resume_from: Optional[Union[str, Path]] = None
+) -> CrawlState
+```
+
+## Parameters
+
+### start_url
+- **Type**: `str`
+- **Required**: Yes
+- **Description**: The starting URL for the crawl. This should be a valid HTTP/HTTPS URL that serves as the entry point for information gathering.
+
+### query
+- **Type**: `str`  
+- **Required**: Yes
+- **Description**: The search query that guides the crawling process. This should contain key terms related to the information you're seeking. The crawler uses this to evaluate relevance and determine which links to follow.
+
+### resume_from
+- **Type**: `Optional[Union[str, Path]]`
+- **Default**: `None`
+- **Description**: Path to a previously saved crawl state file. When provided, the crawler resumes from the saved state instead of starting fresh.
+
+## Return Value
+
+Returns a `CrawlState` object containing:
+
+- **crawled_urls** (`Set[str]`): All URLs that have been crawled
+- **knowledge_base** (`List[CrawlResult]`): Collection of crawled pages with content
+- **pending_links** (`List[Link]`): Links discovered but not yet crawled
+- **metrics** (`Dict[str, float]`): Performance and quality metrics
+- **query** (`str`): The original query
+- Additional statistical information for scoring
+
+## How It Works
+
+The `digest()` method implements an intelligent crawling algorithm:
+
+1. **Initial Crawl**: Starts from the provided URL
+2. **Link Analysis**: Evaluates all discovered links for relevance
+3. **Scoring**: Uses three metrics to assess information sufficiency:
+   - **Coverage**: How well the query terms are covered
+   - **Consistency**: Information coherence across pages
+   - **Saturation**: Diminishing returns detection
+4. **Adaptive Selection**: Chooses the most promising links to follow
+5. **Stopping Decision**: Automatically stops when confidence threshold is reached
+
+## Examples
+
+### Basic Usage
+
+```python
+async with AsyncWebCrawler() as crawler:
+    adaptive = AdaptiveCrawler(crawler)
+    
+    state = await adaptive.digest(
+        start_url="https://docs.python.org/3/",
+        query="async await context managers"
+    )
+    
+    print(f"Crawled {len(state.crawled_urls)} pages")
+    print(f"Confidence: {adaptive.confidence:.0%}")
+```
+
+### With Configuration
+
+```python
+config = AdaptiveConfig(
+    confidence_threshold=0.9,  # Require high confidence
+    max_pages=30,             # Allow more pages
+    top_k_links=3             # Follow top 3 links per page
+)
+
+adaptive = AdaptiveCrawler(crawler, config=config)
+
+state = await adaptive.digest(
+    start_url="https://api.example.com/docs",
+    query="authentication endpoints rate limits"
+)
+```
+
+### Resuming a Previous Crawl
+
+```python
+# First crawl - may be interrupted
+state1 = await adaptive.digest(
+    start_url="https://example.com",
+    query="machine learning algorithms"
+)
+
+# Save state (if not auto-saved)
+state1.save("ml_crawl_state.json")
+
+# Later, resume from saved state
+state2 = await adaptive.digest(
+    start_url="https://example.com",
+    query="machine learning algorithms",
+    resume_from="ml_crawl_state.json"
+)
+```
+
+### With Progress Monitoring
+
+```python
+state = await adaptive.digest(
+    start_url="https://docs.example.com",
+    query="api reference"
+)
+
+# Monitor progress
+print(f"Pages crawled: {len(state.crawled_urls)}")
+print(f"New terms discovered: {state.new_terms_history}")
+print(f"Final confidence: {adaptive.confidence:.2%}")
+
+# View detailed statistics
+adaptive.print_stats(detailed=True)
+```
+
+## Query Best Practices
+
+1. **Be Specific**: Use descriptive terms that appear in target content
+   ```python
+   # Good
+   query = "python async context managers implementation"
+   
+   # Too broad
+   query = "python programming"
+   ```
+
+2. **Include Key Terms**: Add technical terms you expect to find
+   ```python
+   query = "oauth2 jwt refresh tokens authorization"
+   ```
+
+3. **Multiple Concepts**: Combine related concepts for comprehensive coverage
+   ```python
+   query = "rest api pagination sorting filtering"
+   ```
+
+## Performance Considerations
+
+- **Initial URL**: Choose a page with good navigation (e.g., documentation index)
+- **Query Length**: 3-8 terms typically work best
+- **Link Density**: Sites with clear navigation crawl more efficiently
+- **Caching**: Enable caching for repeated crawls of the same domain
+
+## Error Handling
+
+```python
+try:
+    state = await adaptive.digest(
+        start_url="https://example.com",
+        query="search terms"
+    )
+except Exception as e:
+    print(f"Crawl failed: {e}")
+    # State is auto-saved if save_state=True in config
+```
+
+## Stopping Conditions
+
+The crawl stops when any of these conditions are met:
+
+1. **Confidence Threshold**: Reached the configured confidence level
+2. **Page Limit**: Crawled the maximum number of pages
+3. **Diminishing Returns**: Expected information gain below threshold
+4. **No Relevant Links**: No promising links remain to follow
+
+## See Also
+
+- [AdaptiveCrawler Class](adaptive-crawler.md)
+- [Adaptive Crawling Guide](../core/adaptive-crawling.md)
+- [Configuration Options](../core/adaptive-crawling.md#configuration-options)

docs/md_v2/api/parameters.md

@@ -208,6 +208,71 @@ config = CrawlerRunConfig(
 
 See [Virtual Scroll documentation](../../advanced/virtual-scroll.md) for detailed examples.
 
+---
+
+### I) **URL Matching Configuration**
+
+| **Parameter**          | **Type / Default**           | **What It Does**                                                                                                                    |
+|------------------------|------------------------------|-------------------------------------------------------------------------------------------------------------------------------------|
+| **`url_matcher`**      | `UrlMatcher` (None)          | Pattern(s) to match URLs against. Can be: string (glob), function, or list of mixed types. **None means match ALL URLs**         |
+| **`match_mode`**       | `MatchMode` (MatchMode.OR)   | How to combine multiple matchers in a list: `MatchMode.OR` (any match) or `MatchMode.AND` (all must match)                       |
+
+The `url_matcher` parameter enables URL-specific configurations when used with `arun_many()`:
+
+```python
+from crawl4ai import CrawlerRunConfig, MatchMode
+from crawl4ai.processors.pdf import PDFContentScrapingStrategy
+from crawl4ai.extraction_strategy import JsonCssExtractionStrategy
+
+# Simple string pattern (glob-style)
+pdf_config = CrawlerRunConfig(
+    url_matcher="*.pdf",
+    scraping_strategy=PDFContentScrapingStrategy()
+)
+
+# Multiple patterns with OR logic (default)
+blog_config = CrawlerRunConfig(
+    url_matcher=["*/blog/*", "*/article/*", "*/news/*"],
+    match_mode=MatchMode.OR  # Any pattern matches
+)
+
+# Function matcher
+api_config = CrawlerRunConfig(
+    url_matcher=lambda url: 'api' in url or url.endswith('.json'),
+    # Other settings like extraction_strategy
+)
+
+# Mixed: String + Function with AND logic
+complex_config = CrawlerRunConfig(
+    url_matcher=[
+        lambda url: url.startswith('https://'),  # Must be HTTPS
+        "*.org/*",                               # Must be .org domain
+        lambda url: 'docs' in url                # Must contain 'docs'
+    ],
+    match_mode=MatchMode.AND  # ALL conditions must match
+)
+
+# Combined patterns and functions with AND logic
+secure_docs = CrawlerRunConfig(
+    url_matcher=["https://*", lambda url: '.doc' in url],
+    match_mode=MatchMode.AND  # Must be HTTPS AND contain .doc
+)
+
+# Default config - matches ALL URLs
+default_config = CrawlerRunConfig()  # No url_matcher = matches everything
+```
+
+**UrlMatcher Types:**
+- **None (default)**: When `url_matcher` is None or not set, the config matches ALL URLs
+- **String patterns**: Glob-style patterns like `"*.pdf"`, `"*/api/*"`, `"https://*.example.com/*"`
+- **Functions**: `lambda url: bool` - Custom logic for complex matching
+- **Lists**: Mix strings and functions, combined with `MatchMode.OR` or `MatchMode.AND`
+
+**Important Behavior:**
+- When passing a list of configs to `arun_many()`, URLs are matched against each config's `url_matcher` in order. First match wins!
+- If no config matches a URL and there's no default config (one without `url_matcher`), the URL will fail with "No matching configuration found"
+- Always include a default config as the last item if you want to handle all URLs
+
 ---## 2.2 Helper Methods
 
 Both `BrowserConfig` and `CrawlerRunConfig` provide a `clone()` method to create modified copies:
@@ -298,7 +363,7 @@ LLMConfig is useful to pass LLM provider config to strategies and functions that
 ## 3.1 Parameters
 | **Parameter**         | **Type / Default**                     | **What It Does**                                                                                                                     |
 |-----------------------|----------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------|
-| **`provider`**    | `"ollama/llama3","groq/llama3-70b-8192","groq/llama3-8b-8192", "openai/gpt-4o-mini" ,"openai/gpt-4o","openai/o1-mini","openai/o1-preview","openai/o3-mini","openai/o3-mini-high","anthropic/claude-3-haiku-20240307","anthropic/claude-3-opus-20240229","anthropic/claude-3-sonnet-20240229","anthropic/claude-3-5-sonnet-20240620","gemini/gemini-pro","gemini/gemini-1.5-pro","gemini/gemini-2.0-flash","gemini/gemini-2.0-flash-exp","gemini/gemini-2.0-flash-lite-preview-02-05","deepseek/deepseek-chat"`<br/>*(default: `"openai/gpt-4o-mini"`)* | Which LLM provoder to use. 
+| **`provider`**    | `"ollama/llama3","groq/llama3-70b-8192","groq/llama3-8b-8192", "openai/gpt-4o-mini" ,"openai/gpt-4o","openai/o1-mini","openai/o1-preview","openai/o3-mini","openai/o3-mini-high","anthropic/claude-3-haiku-20240307","anthropic/claude-3-opus-20240229","anthropic/claude-3-sonnet-20240229","anthropic/claude-3-5-sonnet-20240620","gemini/gemini-pro","gemini/gemini-1.5-pro","gemini/gemini-2.0-flash","gemini/gemini-2.0-flash-exp","gemini/gemini-2.0-flash-lite-preview-02-05","deepseek/deepseek-chat"`<br/>*(default: `"openai/gpt-4o-mini"`)* | Which LLM provider to use. 
 | **`api_token`**         |1.Optional. When not provided explicitly, api_token will be read from environment variables based on provider. For example: If a gemini model is passed as provider then,`"GEMINI_API_KEY"` will be read from environment variables  <br/> 2. API token of LLM provider <br/> eg: `api_token = "gsk_1ClHGGJ7Lpn4WGybR7vNWGdyb3FY7zXEw3SCiy0BAVM9lL8CQv"` <br/> 3. Environment variable - use with prefix "env:" <br/> eg:`api_token = "env: GROQ_API_KEY"`              | API token to use for the given provider 
 | **`base_url`**         |Optional. Custom API endpoint | If your provider has a custom endpoint
 

docs/md_v2/blog/articles/adaptive-crawling-revolution.md

@@ -0,0 +1,369 @@
+# Adaptive Crawling: Building Dynamic Knowledge That Grows on Demand
+
+*Published on January 29, 2025 • 8 min read*
+
+*By [unclecode](https://x.com/unclecode) • Follow me on [X/Twitter](https://x.com/unclecode) for more web scraping insights*
+
+---
+
+## The Knowledge Capacitor
+
+Imagine a capacitor that stores energy, releasing it precisely when needed. Now imagine that for information. That's Adaptive Crawling—a term I coined to describe a fundamentally different approach to web crawling. Instead of the brute force of traditional deep crawling, we build knowledge dynamically, growing it based on queries and circumstances, like a living organism responding to its environment.
+
+This isn't just another crawling optimization. It's a paradigm shift from "crawl everything, hope for the best" to "crawl intelligently, know when to stop."
+
+## Why I Built This
+
+I've watched too many startups burn through resources with a dangerous misconception: that LLMs make everything efficient. They don't. They make things *possible*, not necessarily *smart*. When you combine brute-force crawling with LLM processing, you're not just wasting time—you're hemorrhaging money on tokens, compute, and opportunity cost.
+
+Consider this reality:
+- **Traditional deep crawling**: 500 pages → 50 useful → $15 in LLM tokens → 2 hours wasted
+- **Adaptive crawling**: 15 pages → 14 useful → $2 in tokens → 10 minutes → **7.5x cost reduction**
+
+But it's not about crawling less. It's about crawling *right*.
+
+## The Information Theory Foundation
+
+<div style="background-color: #1a1a1c; border: 1px solid #3f3f44; padding: 20px; margin: 20px 0;">
+
+### 🧮 **Pure Statistics, No Magic**
+
+My first principle was crucial: start with classic statistical approaches. No embeddings. No LLMs. Just pure information theory:
+
+```python
+# Information gain calculation - the heart of adaptive crawling
+def calculate_information_gain(new_page, knowledge_base):
+    new_terms = extract_terms(new_page) - existing_terms(knowledge_base)
+    overlap = calculate_overlap(new_page, knowledge_base)
+    
+    # High gain = many new terms + low overlap
+    gain = len(new_terms) / (1 + overlap)
+    return gain
+```
+
+This isn't regression to older methods—it's recognition that we've forgotten powerful, efficient solutions in our rush to apply LLMs everywhere.
+
+</div>
+
+## The A* of Web Crawling
+
+Adaptive crawling implements what I call "information scenting"—like A* pathfinding but for knowledge acquisition. Each link is evaluated not randomly, but by its probability of contributing meaningful information toward answering current and future queries.
+
+<div style="display: flex; align-items: center; background-color: #3f3f44; padding: 20px; margin: 20px 0; border-left: 4px solid #09b5a5;">
+<div style="font-size: 48px; margin-right: 20px;">🎯</div>
+<div>
+<strong>The Scenting Algorithm:</strong><br>
+From available links, we select those with highest information gain. It's not about following every path—it's about following the <em>right</em> paths. Like a bloodhound following the strongest scent to its target.
+</div>
+</div>
+
+## The Three Pillars of Intelligence
+
+### 1. Coverage: The Breadth Sensor
+Measures how well your knowledge spans the query space. Not just "do we have pages?" but "do we have the RIGHT pages?"
+
+### 2. Consistency: The Coherence Detector  
+Information from multiple sources should align. When pages agree, confidence rises. When they conflict, we need more data.
+
+### 3. Saturation: The Efficiency Guardian
+The most crucial metric. When new pages stop adding information, we stop crawling. Simple. Powerful. Ignored by everyone else.
+
+## Real Impact: Time, Money, and Sanity
+
+Let me show you what this means for your bottom line:
+
+### Building a Customer Support Knowledge Base
+
+**Traditional Approach:**
+```python
+# Crawl entire documentation site
+results = await crawler.crawl_bfs("https://docs.company.com", max_depth=5)
+# Result: 1,200 pages, 18 hours, $150 in API costs
+# Useful content: ~100 pages scattered throughout
+```
+
+**Adaptive Approach:**
+```python
+# Grow knowledge based on actual support queries
+knowledge = await adaptive.digest(
+    start_url="https://docs.company.com",
+    query="payment processing errors refund policies"
+)
+# Result: 45 pages, 12 minutes, $8 in API costs
+# Useful content: 42 pages, all relevant
+```
+
+**Savings: 93% time reduction, 95% cost reduction, 100% more sanity**
+
+## The Dynamic Growth Pattern
+
+<div style="text-align: center; padding: 40px; background-color: #1a1a1c; border: 1px dashed #3f3f44; margin: 30px 0;">
+<div style="font-size: 24px; color: #09b5a5; margin-bottom: 10px;">
+Knowledge grows like crystals in a supersaturated solution
+</div>
+<div style="color: #a3abba;">
+Add a query (seed), and relevant information crystallizes around it.<br>
+Change the query, and the knowledge structure adapts.
+</div>
+</div>
+
+This is the beauty of adaptive crawling: your knowledge base becomes a living entity that grows based on actual needs, not hypothetical completeness.
+
+## Why "Adaptive"? 
+
+I specifically chose "Adaptive" because it captures the essence: the system adapts to what it finds. Dense technical documentation might need 20 pages for confidence. A simple FAQ might need just 5. The crawler doesn't follow a recipe—it reads the room and adjusts.
+
+This is my term, my concept, and I have extensive plans for its evolution.
+
+## The Progressive Roadmap
+
+This is just the beginning. My roadmap for Adaptive Crawling:
+
+### Phase 1 (Current): Statistical Foundation
+- Pure information theory approach
+- No dependencies on expensive models
+- Proven efficiency gains
+
+### Phase 2 (Now Available): Embedding Enhancement
+- Semantic understanding layered onto statistical base
+- Still efficient, now even smarter
+- Optional, not required
+
+### Phase 3 (Future): LLM Integration
+- LLMs for complex reasoning tasks only
+- Used surgically, not wastefully
+- Always with statistical foundation underneath
+
+## The Efficiency Revolution
+
+<div style="background-color: #1a1a1c; border: 1px solid #3f3f44; padding: 20px; margin: 20px 0;">
+
+### 💰 **The Economics of Intelligence**
+
+For a typical SaaS documentation crawl:
+
+**Traditional Deep Crawling:**
+- Pages crawled: 1,000
+- Useful pages: 80
+- Time spent: 3 hours
+- LLM tokens used: 2.5M
+- Cost: $75
+- Efficiency: 8%
+
+**Adaptive Crawling:**
+- Pages crawled: 95
+- Useful pages: 88
+- Time spent: 15 minutes
+- LLM tokens used: 200K
+- Cost: $6
+- Efficiency: 93%
+
+**That's not optimization. That's transformation.**
+
+</div>
+
+## Missing the Forest for the Trees
+
+The startup world has a dangerous blind spot. We're so enamored with LLMs that we forget: just because you CAN process everything with an LLM doesn't mean you SHOULD. 
+
+Classic NLP and statistical methods can:
+- Filter irrelevant content before it reaches LLMs
+- Identify patterns without expensive inference
+- Make intelligent decisions in microseconds
+- Scale without breaking the bank
+
+Adaptive crawling proves this. It uses battle-tested information theory to make smart decisions BEFORE expensive processing.
+
+## Your Knowledge, On Demand
+
+```python
+# Monday: Customer asks about authentication
+auth_knowledge = await adaptive.digest(
+    "https://docs.api.com",
+    "oauth jwt authentication"
+)
+
+# Tuesday: They ask about rate limiting
+# The crawler adapts, builds on existing knowledge
+rate_limit_knowledge = await adaptive.digest(
+    "https://docs.api.com", 
+    "rate limiting throttling quotas"
+)
+
+# Your knowledge base grows intelligently, not indiscriminately
+```
+
+## The Competitive Edge
+
+Companies using adaptive crawling will have:
+- **90% lower crawling costs**
+- **Knowledge bases that actually answer questions**
+- **Update cycles in minutes, not days**
+- **Happy customers who find answers fast**
+- **Engineers who sleep at night**
+
+Those still using brute force? They'll wonder why their infrastructure costs keep rising while their customers keep complaining.
+
+## The Embedding Evolution (Now Available!)
+
+<div style="background-color: #1a1a1c; border: 1px solid #3f3f44; padding: 20px; margin: 20px 0;">
+
+### 🧠 **Semantic Understanding Without the Cost**
+
+The embedding strategy brings semantic intelligence while maintaining efficiency:
+
+```python
+# Statistical strategy - great for exact terms
+config_statistical = AdaptiveConfig(
+    strategy="statistical"  # Default
+)
+
+# Embedding strategy - understands concepts
+config_embedding = AdaptiveConfig(
+    strategy="embedding",
+    embedding_model="sentence-transformers/all-MiniLM-L6-v2",
+    n_query_variations=10
+)
+```
+
+**The magic**: It automatically expands your query into semantic variations, maps the coverage space, and identifies gaps to fill intelligently.
+
+</div>
+
+### Real-World Comparison
+
+<div style="display: flex; gap: 20px; margin: 20px 0;">
+<div style="flex: 1; background-color: #1a1a1c; border: 1px solid #3f3f44; padding: 20px;">
+
+**Query**: "authentication oauth"
+
+**Statistical Strategy**:
+- Searches for exact terms
+- 12 pages crawled
+- 78% confidence
+- Fast but literal
+
+</div>
+<div style="flex: 1; background-color: #1a1a1c; border: 1px solid #09b5a5; padding: 20px;">
+
+**Embedding Strategy**:
+- Understands "auth", "login", "SSO"
+- 8 pages crawled
+- 92% confidence
+- Semantic comprehension
+
+</div>
+</div>
+
+### Detecting Irrelevance
+
+One killer feature: the embedding strategy knows when to give up:
+
+```python
+# Crawling Python docs with a cooking query
+result = await adaptive.digest(
+    start_url="https://docs.python.org/3/",
+    query="how to make spaghetti carbonara"
+)
+
+# System detects irrelevance and stops
+# Confidence: 5% (below threshold)
+# Pages crawled: 2
+# Stopped reason: "below_minimum_relevance_threshold"
+```
+
+No more crawling hundreds of pages hoping to find something that doesn't exist!
+
+## Try It Yourself
+
+```python
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+
+async with AsyncWebCrawler() as crawler:
+    # Choose your strategy
+    config = AdaptiveConfig(
+        strategy="embedding",  # or "statistical"
+        embedding_min_confidence_threshold=0.1  # Stop if irrelevant
+    )
+    
+    adaptive = AdaptiveCrawler(crawler, config)
+    
+    # Watch intelligence at work
+    result = await adaptive.digest(
+        start_url="https://your-docs.com",
+        query="your users' actual questions"
+    )
+    
+    # See the efficiency
+    adaptive.print_stats()
+    print(f"Found {adaptive.confidence:.0%} of needed information")
+    print(f"In just {len(result.crawled_urls)} pages")
+    print(f"Saving you {1000 - len(result.crawled_urls)} unnecessary crawls")
+```
+
+## A Personal Note
+
+I created Adaptive Crawling because I was tired of watching smart people make inefficient choices. We have incredibly powerful statistical tools that we've forgotten in our rush toward LLMs. This is my attempt to bring balance back to the Force.
+
+This is not just a feature. It's a philosophy: **Grow knowledge on demand. Stop when you have enough. Save time, money, and computational resources for what really matters.**
+
+## The Future is Adaptive
+
+<div style="text-align: center; padding: 40px; background-color: #1a1a1c; border: 1px dashed #3f3f44; margin: 30px 0;">
+<div style="font-size: 24px; color: #09b5a5; margin-bottom: 10px;">
+Traditional Crawling: Drinking from a firehose<br>
+Adaptive Crawling: Sipping exactly what you need
+</div>
+<div style="color: #a3abba;">
+The future of web crawling isn't about processing more data.<br>
+It's about processing the <em>right</em> data.
+</div>
+</div>
+
+Join me in making web crawling intelligent, efficient, and actually useful. Because in the age of information overload, the winners won't be those who collect the most data—they'll be those who collect the *right* data.
+
+---
+
+*Adaptive Crawling is now part of Crawl4AI. [Get started with the documentation](/core/adaptive-crawling/) or [dive into the mathematical framework](https://github.com/unclecode/crawl4ai/blob/main/PROGRESSIVE_CRAWLING.md). For updates on my work in information theory and efficient AI, follow me on [X/Twitter](https://x.com/unclecode).*
+
+<style>
+/* Custom styles for this article */
+.markdown-body pre {
+    background-color: #1e1e1e !important;
+    border: 1px solid #3f3f44;
+}
+
+.markdown-body code {
+    background-color: #3f3f44;
+    color: #50ffff;
+    padding: 2px 6px;
+    border-radius: 3px;
+}
+
+.markdown-body pre code {
+    background-color: transparent;
+    color: #e8e9ed;
+    padding: 0;
+}
+
+.markdown-body blockquote {
+    border-left: 4px solid #09b5a5;
+    background-color: #1a1a1c;
+    padding: 15px 20px;
+    margin: 20px 0;
+}
+
+.markdown-body h2 {
+    color: #50ffff;
+    border-bottom: 1px dashed #3f3f44;
+    padding-bottom: 10px;
+}
+
+.markdown-body h3 {
+    color: #09b5a5;
+}
+
+.markdown-body strong {
+    color: #50ffff;
+}
+</style>

docs/md_v2/blog/index.md

@@ -2,16 +2,46 @@
 
 Welcome to the Crawl4AI blog! Here you'll find detailed release notes, technical insights, and updates about the project. Whether you're looking for the latest improvements or want to dive deep into web crawling techniques, this is the place.
 
+## Featured Articles
+
+### [When to Stop Crawling: The Art of Knowing "Enough"](articles/adaptive-crawling-revolution.md)
+*January 29, 2025*
+
+Traditional crawlers are like tourists with unlimited time—they'll visit every street, every alley, every dead end. But what if your crawler could think like a researcher with a deadline? Discover how Adaptive Crawling revolutionizes web scraping by knowing when to stop. Learn about the three-layer intelligence system that evaluates coverage, consistency, and saturation to build focused knowledge bases instead of endless page collections.
+
+[Read the full article →](articles/adaptive-crawling-revolution.md)
+
+### [The LLM Context Protocol: Why Your AI Assistant Needs Memory, Reasoning, and Examples](articles/llm-context-revolution.md)
+*January 24, 2025*
+
+Ever wondered why your AI coding assistant struggles with your library despite comprehensive documentation? This article introduces the three-dimensional context protocol that transforms how AI understands code. Learn why memory, reasoning, and examples together create wisdom—not just information.
+
+[Read the full article →](articles/llm-context-revolution.md)
+
 ## Latest Release
 
-Here’s the blog index entry for **v0.6.0**, written to match the exact tone and structure of your previous entries:
+### [Crawl4AI v0.7.0 – The Adaptive Intelligence Update](releases/0.7.0.md)
+*January 28, 2025*
+
+Crawl4AI v0.7.0 introduces groundbreaking intelligence features that transform how crawlers understand and adapt to websites. This release brings Adaptive Crawling that learns website patterns, Virtual Scroll support for infinite pages, intelligent Link Preview with 3-layer scoring, and the powerful Async URL Seeder for massive URL discovery.
+
+Key highlights:
+- **Adaptive Crawling**: Crawlers that learn and adapt to website structures automatically
+- **Virtual Scroll Support**: Complete content extraction from modern infinite scroll pages  
+- **Link Preview**: 3-layer scoring system for intelligent link prioritization
+- **Async URL Seeder**: Discover thousands of URLs in seconds with smart filtering
+- **Performance Boost**: Up to 3x faster with optimized resource handling
+
+[Read full release notes →](releases/0.7.0.md)
 
 ---
 
-### [Crawl4AI v0.6.0 – World-Aware Crawling, Pre-Warmed Browsers, and the MCP API](releases/0.6.0.md)
-*April 23, 2025*
+## Previous Releases
 
-Crawl4AI v0.6.0 is our most powerful release yet. This update brings major architectural upgrades including world-aware crawling (set geolocation, locale, and timezone), real-time traffic capture, and a memory-efficient crawler pool with pre-warmed pages.  
+### [Crawl4AI v0.6.0 – World-Aware Crawling, Pre-Warmed Browsers, and the MCP API](releases/0.6.0.md)
+*December 23, 2024*
+
+Crawl4AI v0.6.0 brought major architectural upgrades including world-aware crawling (set geolocation, locale, and timezone), real-time traffic capture, and a memory-efficient crawler pool with pre-warmed pages.  
 
 The Docker server now exposes a full-featured MCP socket + SSE interface, supports streaming, and comes with a new Playground UI. Plus, table extraction is now native, and the new stress-test framework supports crawling 1,000+ URLs.  
 
@@ -29,8 +59,6 @@ Other key changes:
 
 ---
 
-Let me know if you want me to auto-update the actual file or just paste this into the markdown.
-
 ### [Crawl4AI v0.5.0: Deep Crawling, Scalability, and a New CLI!](releases/0.5.0.md)
 
 My dear friends and crawlers, there you go, this is the release of Crawl4AI v0.5.0! This release brings a wealth of new features, performance improvements, and a more streamlined developer experience.  Here's a breakdown of what's new:
@@ -124,5 +152,4 @@ Curious about how Crawl4AI has evolved? Check out our [complete changelog](https
 
 - Star us on [GitHub](https://github.com/unclecode/crawl4ai)
 - Follow [@unclecode](https://twitter.com/unclecode) on Twitter
-- Join our community discussions on GitHub
-
+- Join our community discussions on GitHub

docs/md_v2/blog/index.md.bak

@@ -0,0 +1,144 @@
+# Crawl4AI Blog
+
+Welcome to the Crawl4AI blog! Here you'll find detailed release notes, technical insights, and updates about the project. Whether you're looking for the latest improvements or want to dive deep into web crawling techniques, this is the place.
+
+## Featured Articles
+
+### [When to Stop Crawling: The Art of Knowing "Enough"](articles/adaptive-crawling-revolution.md)
+*January 29, 2025*
+
+Traditional crawlers are like tourists with unlimited time—they'll visit every street, every alley, every dead end. But what if your crawler could think like a researcher with a deadline? Discover how Adaptive Crawling revolutionizes web scraping by knowing when to stop. Learn about the three-layer intelligence system that evaluates coverage, consistency, and saturation to build focused knowledge bases instead of endless page collections.
+
+[Read the full article →](articles/adaptive-crawling-revolution.md)
+
+### [The LLM Context Protocol: Why Your AI Assistant Needs Memory, Reasoning, and Examples](articles/llm-context-revolution.md)
+*January 24, 2025*
+
+Ever wondered why your AI coding assistant struggles with your library despite comprehensive documentation? This article introduces the three-dimensional context protocol that transforms how AI understands code. Learn why memory, reasoning, and examples together create wisdom—not just information.
+
+[Read the full article →](articles/llm-context-revolution.md)
+
+## Latest Release
+
+Here’s the blog index entry for **v0.6.0**, written to match the exact tone and structure of your previous entries:
+
+---
+
+### [Crawl4AI v0.6.0 – World-Aware Crawling, Pre-Warmed Browsers, and the MCP API](releases/0.6.0.md)
+*April 23, 2025*
+
+Crawl4AI v0.6.0 is our most powerful release yet. This update brings major architectural upgrades including world-aware crawling (set geolocation, locale, and timezone), real-time traffic capture, and a memory-efficient crawler pool with pre-warmed pages.  
+
+The Docker server now exposes a full-featured MCP socket + SSE interface, supports streaming, and comes with a new Playground UI. Plus, table extraction is now native, and the new stress-test framework supports crawling 1,000+ URLs.  
+
+Other key changes:  
+
+*   Native support for `result.media["tables"]` to export DataFrames  
+* Full network + console logs and MHTML snapshot per crawl  
+* Browser pooling and pre-warming for faster cold starts  
+* New streaming endpoints via MCP API and Playground  
+* Robots.txt support, proxy rotation, and improved session handling  
+* Deprecated old markdown names, legacy modules cleaned up  
+* Massive repo cleanup: ~36K insertions, ~5K deletions across 121 files
+
+[Read full release notes →](releases/0.6.0.md)
+
+---
+
+Let me know if you want me to auto-update the actual file or just paste this into the markdown.
+
+### [Crawl4AI v0.5.0: Deep Crawling, Scalability, and a New CLI!](releases/0.5.0.md)
+
+My dear friends and crawlers, there you go, this is the release of Crawl4AI v0.5.0! This release brings a wealth of new features, performance improvements, and a more streamlined developer experience.  Here's a breakdown of what's new:
+
+**Major New Features:**
+
+*   **Deep Crawling:** Explore entire websites with configurable strategies (BFS, DFS, Best-First).  Define custom filters and URL scoring for targeted crawls.
+*   **Memory-Adaptive Dispatcher:**  Handle large-scale crawls with ease!  Our new dispatcher dynamically adjusts concurrency based on available memory and includes built-in rate limiting.
+*   **Multiple Crawler Strategies:** Choose between the full-featured Playwright browser-based crawler or a new, *much* faster HTTP-only crawler for simpler tasks.
+*   **Docker Deployment:**  Deploy Crawl4AI as a scalable, self-contained service with built-in API endpoints and optional JWT authentication.
+*   **Command-Line Interface (CLI):**  Interact with Crawl4AI directly from your terminal.  Crawl, configure, and extract data with simple commands.
+*   **LLM Configuration (`LLMConfig`):** A new, unified way to configure LLM providers (OpenAI, Anthropic, Ollama, etc.) for extraction, filtering, and schema generation.  Simplifies API key management and switching between models.
+
+**Minor Updates & Improvements:**
+
+*   **LXML Scraping Mode:** Faster HTML parsing with `LXMLWebScrapingStrategy`.
+*   **Proxy Rotation:** Added `ProxyRotationStrategy` with a `RoundRobinProxyStrategy` implementation.
+*   **PDF Processing:** Extract text, images, and metadata from PDF files.
+*   **URL Redirection Tracking:**  Automatically follows and records redirects.
+*   **Robots.txt Compliance:**  Optionally respect website crawling rules.
+*   **LLM-Powered Schema Generation:**  Automatically create extraction schemas using an LLM.
+*   **`LLMContentFilter`:** Generate high-quality, focused markdown using an LLM.
+*   **Improved Error Handling & Stability:** Numerous bug fixes and performance enhancements.
+*   **Enhanced Documentation:**  Updated guides and examples.
+
+**Breaking Changes & Migration:**
+
+This release includes several breaking changes to improve the library's structure and consistency.  Here's what you need to know:
+
+*   **`arun_many()` Behavior:** Now uses the `MemoryAdaptiveDispatcher` by default.  The return type depends on the `stream` parameter in `CrawlerRunConfig`.  Adjust code that relied on unbounded concurrency.
+*   **`max_depth` Location:** Moved to `CrawlerRunConfig` and now controls *crawl depth*.
+*   **Deep Crawling Imports:**  Import `DeepCrawlStrategy` and related classes from `crawl4ai.deep_crawling`.
+*   **`BrowserContext` API:**  Updated; the old `get_context` method is deprecated.
+*   **Optional Model Fields:** Many data model fields are now optional.  Handle potential `None` values.
+*   **`ScrapingMode` Enum:** Replaced with strategy pattern (`WebScrapingStrategy`, `LXMLWebScrapingStrategy`).
+*   **`content_filter` Parameter:** Removed from `CrawlerRunConfig`. Use extraction strategies or markdown generators with filters.
+*   **Removed Functionality:** The synchronous `WebCrawler`, the old CLI, and docs management tools have been removed.
+*   **Docker:**  Significant changes to deployment.  See the [Docker documentation](../deploy/docker/README.md).
+*   **`ssl_certificate.json`:** This file has been removed.
+* **Config**: FastFilterChain has been replaced with FilterChain
+* **Deep-Crawl**: DeepCrawlStrategy.arun now returns Union[CrawlResultT, List[CrawlResultT], AsyncGenerator[CrawlResultT, None]]
+* **Proxy**: Removed synchronous WebCrawler support and related rate limiting configurations
+*   **LLM Parameters:** Use the new `LLMConfig` object instead of passing `provider`, `api_token`, `base_url`, and `api_base` directly to `LLMExtractionStrategy` and `LLMContentFilter`.
+
+**In short:** Update imports, adjust `arun_many()` usage, check for optional fields, and review the Docker deployment guide.
+
+## License Change
+
+Crawl4AI v0.5.0 updates the license to Apache 2.0 *with a required attribution clause*.  This means you are free to use, modify, and distribute Crawl4AI (even commercially), but you *must* clearly attribute the project in any public use or distribution.  See the updated `LICENSE` file for the full legal text and specific requirements.
+
+**Get Started:**
+
+*   **Installation:** `pip install "crawl4ai[all]"` (or use the Docker image)
+*   **Documentation:** [https://docs.crawl4ai.com](https://docs.crawl4ai.com)
+*   **GitHub:** [https://github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+
+I'm very excited to see what you build with Crawl4AI v0.5.0!
+
+---
+
+### [0.4.2 - Configurable Crawlers, Session Management, and Smarter Screenshots](releases/0.4.2.md)
+*December 12, 2024*
+
+The 0.4.2 update brings massive improvements to configuration, making crawlers and browsers easier to manage with dedicated objects. You can now import/export local storage for seamless session management. Plus, long-page screenshots are faster and cleaner, and full-page PDF exports are now possible. Check out all the new features to make your crawling experience even smoother.
+
+[Read full release notes →](releases/0.4.2.md)
+
+---
+
+### [0.4.1 - Smarter Crawling with Lazy-Load Handling, Text-Only Mode, and More](releases/0.4.1.md)
+*December 8, 2024*
+
+This release brings major improvements to handling lazy-loaded images, a blazing-fast Text-Only Mode, full-page scanning for infinite scrolls, dynamic viewport adjustments, and session reuse for efficient crawling. If you're looking to improve speed, reliability, or handle dynamic content with ease, this update has you covered.
+
+[Read full release notes →](releases/0.4.1.md)
+
+---
+
+### [0.4.0 - Major Content Filtering Update](releases/0.4.0.md)
+*December 1, 2024*
+
+Introduced significant improvements to content filtering, multi-threaded environment handling, and user-agent generation. This release features the new PruningContentFilter, enhanced thread safety, and improved test coverage.
+
+[Read full release notes →](releases/0.4.0.md)
+
+## Project History
+
+Curious about how Crawl4AI has evolved? Check out our [complete changelog](https://github.com/unclecode/crawl4ai/blob/main/CHANGELOG.md) for a detailed history of all versions and updates.
+
+## Stay Updated
+
+- Star us on [GitHub](https://github.com/unclecode/crawl4ai)
+- Follow [@unclecode](https://twitter.com/unclecode) on Twitter
+- Join our community discussions on GitHub
+

docs/md_v2/blog/releases/0.7.0.md

@@ -0,0 +1,343 @@
+# 🚀 Crawl4AI v0.7.0: The Adaptive Intelligence Update
+
+*January 28, 2025 • 10 min read*
+
+---
+
+Today I'm releasing Crawl4AI v0.7.0—the Adaptive Intelligence Update. This release introduces fundamental improvements in how Crawl4AI handles modern web complexity through adaptive learning, intelligent content discovery, and advanced extraction capabilities.
+
+## 🎯 What's New at a Glance
+
+- **Adaptive Crawling**: Your crawler now learns and adapts to website patterns
+- **Virtual Scroll Support**: Complete content extraction from infinite scroll pages
+- **Link Preview with Intelligent Scoring**: Intelligent link analysis and prioritization
+- **Async URL Seeder**: Discover thousands of URLs in seconds with intelligent filtering
+- **Performance Optimizations**: Significant speed and memory improvements
+
+## 🧠 Adaptive Crawling: Intelligence Through Pattern Learning
+
+**The Problem:** Websites change. Class names shift. IDs disappear. Your carefully crafted selectors break at 3 AM, and you wake up to empty datasets and angry stakeholders.
+
+**My Solution:** I implemented an adaptive learning system that observes patterns, builds confidence scores, and adjusts extraction strategies on the fly. It's like having a junior developer who gets better at their job with every page they scrape.
+
+### Technical Deep-Dive
+
+The Adaptive Crawler maintains a persistent state for each domain, tracking:
+- Pattern success rates
+- Selector stability over time  
+- Content structure variations
+- Extraction confidence scores
+
+```python
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+import asyncio
+
+async def main():
+    
+    # Configure adaptive crawler
+    config = AdaptiveConfig(
+        strategy="statistical",  # or "embedding" for semantic understanding
+        max_pages=10,
+        confidence_threshold=0.7,  # Stop at 70% confidence
+        top_k_links=3,  # Follow top 3 links per page
+        min_gain_threshold=0.05  # Need 5% information gain to continue
+    )
+    
+    async with AsyncWebCrawler(verbose=False) as crawler:
+        adaptive = AdaptiveCrawler(crawler, config)
+        
+        print("Starting adaptive crawl about Python decorators...")
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/glossary.html",
+            query="python decorators functions wrapping"
+        )
+        
+        print(f"\n✅ Crawling Complete!")
+        print(f"• Confidence Level: {adaptive.confidence:.0%}")
+        print(f"• Pages Crawled: {len(result.crawled_urls)}")
+        print(f"• Knowledge Base: {len(adaptive.state.knowledge_base)} documents")
+        
+        # Get most relevant content
+        relevant = adaptive.get_relevant_content(top_k=3)
+        print(f"\nMost Relevant Pages:")
+        for i, page in enumerate(relevant, 1):
+            print(f"{i}. {page['url']} (relevance: {page['score']:.2%})")
+
+asyncio.run(main())
+```
+
+**Expected Real-World Impact:**
+- **News Aggregation**: Maintain 95%+ extraction accuracy even as news sites update their templates
+- **E-commerce Monitoring**: Track product changes across hundreds of stores without constant maintenance
+- **Research Data Collection**: Build robust academic datasets that survive website redesigns
+- **Reduced Maintenance**: Cut selector update time by 80% for frequently-changing sites
+
+## 🌊 Virtual Scroll: Complete Content Capture
+
+**The Problem:** Modern web apps only render what's visible. Scroll down, new content appears, old content vanishes into the void. Traditional crawlers capture that first viewport and miss 90% of the content. It's like reading only the first page of every book.
+
+**My Solution:** I built Virtual Scroll support that mimics human browsing behavior, capturing content as it loads and preserving it before the browser's garbage collector strikes.
+
+### Implementation Details
+
+```python
+from crawl4ai import VirtualScrollConfig
+
+# For social media feeds (Twitter/X style)
+twitter_config = VirtualScrollConfig(
+    container_selector="[data-testid='primaryColumn']",
+    scroll_count=20,                    # Number of scrolls
+    scroll_by="container_height",       # Smart scrolling by container size
+    wait_after_scroll=1.0              # Let content load
+)
+
+# For e-commerce product grids (Instagram style)
+grid_config = VirtualScrollConfig(
+    container_selector="main .product-grid",
+    scroll_count=30,
+    scroll_by=800,                     # Fixed pixel scrolling
+    wait_after_scroll=1.5              # Images need time
+)
+
+# For news feeds with lazy loading
+news_config = VirtualScrollConfig(
+    container_selector=".article-feed",
+    scroll_count=50,
+    scroll_by="page_height",           # Viewport-based scrolling
+    wait_after_scroll=0.5              # Wait for content to load
+)
+
+# Use it in your crawl
+async with AsyncWebCrawler() as crawler:
+    result = await crawler.arun(
+        "https://twitter.com/trending",
+        config=CrawlerRunConfig(
+            virtual_scroll_config=twitter_config,
+            # Combine with other features
+            extraction_strategy=JsonCssExtractionStrategy({
+                "tweets": {
+                    "selector": "[data-testid='tweet']",
+                    "fields": {
+                        "text": {"selector": "[data-testid='tweetText']", "type": "text"},
+                        "likes": {"selector": "[data-testid='like']", "type": "text"}
+                    }
+                }
+            })
+        )
+    )
+    
+    print(f"Captured {len(result.extracted_content['tweets'])} tweets")
+```
+
+**Key Capabilities:**
+- **DOM Recycling Awareness**: Detects and handles virtual DOM element recycling
+- **Smart Scroll Physics**: Three modes - container height, page height, or fixed pixels
+- **Content Preservation**: Captures content before it's destroyed
+- **Intelligent Stopping**: Stops when no new content appears
+- **Memory Efficient**: Streams content instead of holding everything in memory
+
+**Expected Real-World Impact:**
+- **Social Media Analysis**: Capture entire Twitter threads with hundreds of replies, not just top 10
+- **E-commerce Scraping**: Extract 500+ products from infinite scroll catalogs vs. 20-50 with traditional methods  
+- **News Aggregation**: Get all articles from modern news sites, not just above-the-fold content
+- **Research Applications**: Complete data extraction from academic databases using virtual pagination
+
+## 🔗 Link Preview: Intelligent Link Analysis and Scoring
+
+**The Problem:** You crawl a page and get 200 links. Which ones matter? Which lead to the content you actually want? Traditional crawlers force you to follow everything or build complex filters.
+
+**My Solution:** I implemented a three-layer scoring system that analyzes links like a human would—considering their position, context, and relevance to your goals.
+
+### Intelligent Link Analysis and Scoring
+
+```python
+import asyncio
+from crawl4ai import CrawlerRunConfig, CacheMode, AsyncWebCrawler
+from crawl4ai.adaptive_crawler import LinkPreviewConfig
+
+async def main():
+    # Configure intelligent link analysis
+    link_config = LinkPreviewConfig(
+        include_internal=True,
+        include_external=False,
+        max_links=10,
+        concurrency=5,
+        query="python tutorial",  # For contextual scoring
+        score_threshold=0.3,
+        verbose=True
+    )
+    # Use in your crawl
+    async with AsyncWebCrawler() as crawler:
+        result = await crawler.arun(
+            "https://www.geeksforgeeks.org/",
+            config=CrawlerRunConfig(
+                link_preview_config=link_config,
+                score_links=True,  # Enable intrinsic scoring
+                cache_mode=CacheMode.BYPASS
+            )
+        )
+
+        # Access scored and sorted links
+        if result.success and result.links:
+            for link in result.links.get("internal", []):
+                text = link.get('text', 'No text')[:40]
+                print(
+                    text,
+                    f"{link.get('intrinsic_score', 0):.1f}/10" if link.get('intrinsic_score') is not None else "0.0/10",
+                    f"{link.get('contextual_score', 0):.2f}/1" if link.get('contextual_score') is not None else "0.00/1",
+                    f"{link.get('total_score', 0):.3f}" if link.get('total_score') is not None else "0.000"
+                )
+
+asyncio.run(main())
+```
+
+**Scoring Components:**
+
+1. **Intrinsic Score**: Based on link quality indicators
+   - Position on page (navigation, content, footer)
+   - Link attributes (rel, title, class names)
+   - Anchor text quality and length
+   - URL structure and depth
+
+2. **Contextual Score**: Relevance to your query using BM25 algorithm
+   - Keyword matching in link text and title
+   - Meta description analysis
+   - Content preview scoring
+
+3. **Total Score**: Combined score for final ranking
+
+**Expected Real-World Impact:**
+- **Research Efficiency**: Find relevant papers 10x faster by following only high-score links
+- **Competitive Analysis**: Automatically identify important pages on competitor sites
+- **Content Discovery**: Build topic-focused crawlers that stay on track
+- **SEO Audits**: Identify and prioritize high-value internal linking opportunities
+
+## 🎣 Async URL Seeder: Automated URL Discovery at Scale
+
+**The Problem:** You want to crawl an entire domain but only have the homepage. Or worse, you want specific content types across thousands of pages. Manual URL discovery? That's a job for machines, not humans.
+
+**My Solution:** I built Async URL Seeder—a turbocharged URL discovery engine that combines multiple sources with intelligent filtering and relevance scoring.
+
+### Technical Architecture
+
+```python
+import asyncio
+from crawl4ai import AsyncUrlSeeder, SeedingConfig
+
+async def main():
+    async with AsyncUrlSeeder() as seeder:
+        # Discover Python tutorial URLs
+        config = SeedingConfig(
+            source="sitemap",  # Use sitemap
+            pattern="*python*",  # URL pattern filter
+            extract_head=True,  # Get metadata
+            query="python tutorial",  # For relevance scoring
+            scoring_method="bm25",
+            score_threshold=0.2,
+            max_urls=10
+        )
+        
+        print("Discovering Python async tutorial URLs...")
+        urls = await seeder.urls("https://www.geeksforgeeks.org/", config)
+        
+        print(f"\n✅ Found {len(urls)} relevant URLs:")
+        for i, url_info in enumerate(urls[:5], 1):
+            print(f"\n{i}. {url_info['url']}")
+            if url_info.get('relevance_score'):
+                print(f"   Relevance: {url_info['relevance_score']:.3f}")
+            if url_info.get('head_data', {}).get('title'):
+                print(f"   Title: {url_info['head_data']['title'][:60]}...")
+
+asyncio.run(main())
+```
+
+**Discovery Methods:**
+- **Sitemap Mining**: Parses robots.txt and all linked sitemaps
+- **Common Crawl**: Queries the Common Crawl index for historical URLs
+- **Intelligent Crawling**: Follows links with smart depth control
+- **Pattern Analysis**: Learns URL structures and generates variations
+
+**Expected Real-World Impact:**
+- **Migration Projects**: Discover 10,000+ URLs from legacy sites in under 60 seconds
+- **Market Research**: Map entire competitor ecosystems automatically
+- **Academic Research**: Build comprehensive datasets without manual URL collection
+- **SEO Audits**: Find every indexable page with content scoring
+- **Content Archival**: Ensure no content is left behind during site migrations
+
+## ⚡ Performance Optimizations
+
+This release includes significant performance improvements through optimized resource handling, better concurrency management, and reduced memory footprint.
+
+### What We Optimized
+
+```python
+# Optimized crawling with v0.7.0 improvements
+results = []
+for url in urls:
+    result = await crawler.arun(
+        url,
+        config=CrawlerRunConfig(
+            # Performance optimizations
+            wait_until="domcontentloaded",  # Faster than networkidle
+            cache_mode=CacheMode.ENABLED    # Enable caching
+        )
+    )
+    results.append(result)
+```
+
+**Performance Gains:**
+- **Startup Time**: 70% faster browser initialization
+- **Page Loading**: 40% reduction with smart resource blocking
+- **Extraction**: 3x faster with compiled CSS selectors
+- **Memory Usage**: 60% reduction with streaming processing
+- **Concurrent Crawls**: Handle 5x more parallel requests
+
+
+## 🔧 Important Changes
+
+### Breaking Changes
+- `link_extractor` renamed to `link_preview` (better reflects functionality)
+- Minimum Python version now 3.9
+- `CrawlerConfig` split into `CrawlerRunConfig` and `BrowserConfig`
+
+### Migration Guide
+```python
+# Old (v0.6.x)
+from crawl4ai import CrawlerConfig
+config = CrawlerConfig(timeout=30000)
+
+# New (v0.7.0)
+from crawl4ai import CrawlerRunConfig, BrowserConfig
+browser_config = BrowserConfig(timeout=30000)
+run_config = CrawlerRunConfig(cache_mode=CacheMode.BYPASS)
+```
+
+## 🤖 Coming Soon: Intelligent Web Automation
+
+I'm currently working on bringing advanced automation capabilities to Crawl4AI. This includes:
+
+- **Crawl Agents**: Autonomous crawlers that understand your goals and adapt their strategies
+- **Auto JS Generation**: Automatic JavaScript code generation for complex interactions
+- **Smart Form Handling**: Intelligent form detection and filling
+- **Context-Aware Actions**: Crawlers that understand page context and make decisions
+
+These features are under active development and will revolutionize how we approach web automation. Stay tuned!
+
+## 🚀 Get Started
+
+```bash
+pip install crawl4ai==0.7.0
+```
+
+Check out the [updated documentation](https://docs.crawl4ai.com).
+
+Questions? Issues? I'm always listening:
+- GitHub: [github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+- Discord: [discord.gg/crawl4ai](https://discord.gg/jP8KfhDhyN)
+- Twitter: [@unclecode](https://x.com/unclecode)
+
+Happy crawling! 🕷️
+
+---
+
+*P.S. If you're using Crawl4AI in production, I'd love to hear about it. Your use cases inspire the next features.*

docs/md_v2/blog/releases/0.7.1.md

@@ -0,0 +1,43 @@
+# 🛠️ Crawl4AI v0.7.1: Minor Cleanup Update
+
+*July 17, 2025 • 2 min read*
+
+---
+
+A small maintenance release that removes unused code and improves documentation.
+
+## 🎯 What's Changed
+
+- **Removed unused StealthConfig** from `crawl4ai/browser_manager.py`
+- **Updated documentation** with better examples and parameter explanations
+- **Fixed virtual scroll configuration** examples in docs
+
+## 🧹 Code Cleanup
+
+Removed unused `StealthConfig` import and configuration that wasn't being used anywhere in the codebase. The project uses its own custom stealth implementation through JavaScript injection instead.
+
+```python
+# Removed unused code:
+from playwright_stealth import StealthConfig
+stealth_config = StealthConfig(...)  # This was never used
+```
+
+## 📖 Documentation Updates
+
+- Fixed adaptive crawling parameter examples
+- Updated session management documentation
+- Corrected virtual scroll configuration examples
+
+## 🚀 Installation
+
+```bash
+pip install crawl4ai==0.7.1
+```
+
+No breaking changes - upgrade directly from v0.7.0.
+
+---
+
+Questions? Issues? 
+- GitHub: [github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+- Discord: [discord.gg/crawl4ai](https://discord.gg/jP8KfhDhyN)

docs/md_v2/blog/releases/0.7.2.md

@@ -0,0 +1,98 @@
+# 🚀 Crawl4AI v0.7.2: CI/CD & Dependency Optimization Update
+
+*July 25, 2025 • 3 min read*
+
+---
+
+This release introduces automated CI/CD pipelines for seamless releases and optimizes dependencies for a lighter, more efficient package.
+
+## 🎯 What's New
+
+### 🔄 Automated Release Pipeline
+- **GitHub Actions CI/CD**: Automated PyPI and Docker Hub releases on tag push
+- **Multi-platform Docker images**: Support for both AMD64 and ARM64 architectures
+- **Version consistency checks**: Ensures tag, package, and Docker versions align
+- **Automated release notes**: GitHub releases created automatically
+
+### 📦 Dependency Optimization
+- **Moved sentence-transformers to optional dependencies**: Significantly reduces default installation size
+- **Lighter Docker images**: Optimized Dockerfile for faster builds and smaller images
+- **Better dependency management**: Core vs. optional dependencies clearly separated
+
+## 🏗️ CI/CD Pipeline
+
+The new automated release process ensures consistent, reliable releases:
+
+```yaml
+# Trigger releases with a simple tag
+git tag v0.7.2
+git push origin v0.7.2
+
+# Automatically:
+# ✅ Validates version consistency
+# ✅ Builds and publishes to PyPI
+# ✅ Builds multi-platform Docker images
+# ✅ Pushes to Docker Hub with proper tags
+# ✅ Creates GitHub release
+```
+
+## 💾 Lighter Installation
+
+Default installation is now significantly smaller:
+
+```bash
+# Core installation (smaller, faster)
+pip install crawl4ai==0.7.2
+
+# With ML features (includes sentence-transformers)
+pip install crawl4ai[transformer]==0.7.2
+
+# Full installation
+pip install crawl4ai[all]==0.7.2
+```
+
+## 🐳 Docker Improvements
+
+Enhanced Docker support with multi-platform images:
+
+```bash
+# Pull the latest version
+docker pull unclecode/crawl4ai:0.7.2
+docker pull unclecode/crawl4ai:latest
+
+# Available tags:
+# - unclecode/crawl4ai:0.7.2 (specific version)
+# - unclecode/crawl4ai:0.7 (minor version)
+# - unclecode/crawl4ai:0 (major version)
+# - unclecode/crawl4ai:latest
+```
+
+## 🔧 Technical Details
+
+### Dependency Changes
+- `sentence-transformers` moved from required to optional dependencies
+- Reduces default installation by ~500MB
+- No impact on functionality when transformer features aren't needed
+
+### CI/CD Configuration
+- GitHub Actions workflows for automated releases
+- Version validation before publishing
+- Parallel PyPI and Docker Hub deployments
+- Automatic tagging strategy for Docker images
+
+## 🚀 Installation
+
+```bash
+pip install crawl4ai==0.7.2
+```
+
+No breaking changes - direct upgrade from v0.7.0 or v0.7.1.
+
+---
+
+Questions? Issues? 
+- GitHub: [github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+- Discord: [discord.gg/crawl4ai](https://discord.gg/jP8KfhDhyN)
+- Twitter: [@unclecode](https://x.com/unclecode)
+
+*P.S. The new CI/CD pipeline will make future releases faster and more reliable. Thanks for your patience as we improve our release process!*

docs/md_v2/core/adaptive-crawling.md

@@ -0,0 +1,347 @@
+# Adaptive Web Crawling
+
+## Introduction
+
+Traditional web crawlers follow predetermined patterns, crawling pages blindly without knowing when they've gathered enough information. **Adaptive Crawling** changes this paradigm by introducing intelligence into the crawling process.
+
+Think of it like research: when you're looking for information, you don't read every book in the library. You stop when you've found sufficient information to answer your question. That's exactly what Adaptive Crawling does for web scraping.
+
+## Key Concepts
+
+### The Problem It Solves
+
+When crawling websites for specific information, you face two challenges:
+1. **Under-crawling**: Stopping too early and missing crucial information
+2. **Over-crawling**: Wasting resources by crawling irrelevant pages
+
+Adaptive Crawling solves both by using a three-layer scoring system that determines when you have "enough" information.
+
+### How It Works
+
+The AdaptiveCrawler uses three metrics to measure information sufficiency:
+
+- **Coverage**: How well your collected pages cover the query terms
+- **Consistency**: Whether the information is coherent across pages  
+- **Saturation**: Detecting when new pages aren't adding new information
+
+When these metrics indicate sufficient information has been gathered, crawling stops automatically.
+
+## Quick Start
+
+### Basic Usage
+
+```python
+from crawl4ai import AsyncWebCrawler, AdaptiveCrawler
+
+async def main():
+    async with AsyncWebCrawler() as crawler:
+        # Create an adaptive crawler (config is optional)
+        adaptive = AdaptiveCrawler(crawler)
+        
+        # Start crawling with a query
+        result = await adaptive.digest(
+            start_url="https://docs.python.org/3/",
+            query="async context managers"
+        )
+        
+        # View statistics
+        adaptive.print_stats()
+        
+        # Get the most relevant content
+        relevant_pages = adaptive.get_relevant_content(top_k=5)
+        for page in relevant_pages:
+            print(f"- {page['url']} (score: {page['score']:.2f})")
+```
+
+### Configuration Options
+
+```python
+from crawl4ai import AdaptiveConfig
+
+config = AdaptiveConfig(
+    confidence_threshold=0.8,    # Stop when 80% confident (default: 0.7)
+    max_pages=30,               # Maximum pages to crawl (default: 20)
+    top_k_links=5,              # Links to follow per page (default: 3)
+    min_gain_threshold=0.05     # Minimum expected gain to continue (default: 0.1)
+)
+
+adaptive = AdaptiveCrawler(crawler, config)
+```
+
+## Crawling Strategies
+
+Adaptive Crawling supports two distinct strategies for determining information sufficiency:
+
+### Statistical Strategy (Default)
+
+The statistical strategy uses pure information theory and term-based analysis:
+
+- **Fast and efficient** - No API calls or model loading
+- **Term-based coverage** - Analyzes query term presence and distribution
+- **No external dependencies** - Works offline
+- **Best for**: Well-defined queries with specific terminology
+
+```python
+# Default configuration uses statistical strategy
+config = AdaptiveConfig(
+    strategy="statistical",  # This is the default
+    confidence_threshold=0.8
+)
+```
+
+### Embedding Strategy
+
+The embedding strategy uses semantic embeddings for deeper understanding:
+
+- **Semantic understanding** - Captures meaning beyond exact term matches
+- **Query expansion** - Automatically generates query variations
+- **Gap-driven selection** - Identifies semantic gaps in knowledge
+- **Validation-based stopping** - Uses held-out queries to validate coverage
+- **Best for**: Complex queries, ambiguous topics, conceptual understanding
+
+```python
+# Configure embedding strategy
+config = AdaptiveConfig(
+    strategy="embedding",
+    embedding_model="sentence-transformers/all-MiniLM-L6-v2",  # Default
+    n_query_variations=10,  # Generate 10 query variations
+    embedding_min_confidence_threshold=0.1  # Stop if completely irrelevant
+)
+
+# With custom embedding provider (e.g., OpenAI)
+config = AdaptiveConfig(
+    strategy="embedding",
+    embedding_llm_config={
+        'provider': 'openai/text-embedding-3-small',
+        'api_token': 'your-api-key'
+    }
+)
+```
+
+### Strategy Comparison
+
+| Feature | Statistical | Embedding |
+|---------|------------|-----------|
+| **Speed** | Very fast | Moderate (API calls) |
+| **Cost** | Free | Depends on provider |
+| **Accuracy** | Good for exact terms | Excellent for concepts |
+| **Dependencies** | None | Embedding model/API |
+| **Query Understanding** | Literal | Semantic |
+| **Best Use Case** | Technical docs, specific terms | Research, broad topics |
+
+### Embedding Strategy Configuration
+
+The embedding strategy offers fine-tuned control through several parameters:
+
+```python
+config = AdaptiveConfig(
+    strategy="embedding",
+    
+    # Model configuration
+    embedding_model="sentence-transformers/all-MiniLM-L6-v2",
+    embedding_llm_config=None,  # Use for API-based embeddings
+    
+    # Query expansion
+    n_query_variations=10,  # Number of query variations to generate
+    
+    # Coverage parameters
+    embedding_coverage_radius=0.2,  # Distance threshold for coverage
+    embedding_k_exp=3.0,  # Exponential decay factor (higher = stricter)
+    
+    # Stopping criteria
+    embedding_min_relative_improvement=0.1,  # Min improvement to continue
+    embedding_validation_min_score=0.3,  # Min validation score
+    embedding_min_confidence_threshold=0.1,  # Below this = irrelevant
+    
+    # Link selection
+    embedding_overlap_threshold=0.85,  # Similarity for deduplication
+    
+    # Display confidence mapping
+    embedding_quality_min_confidence=0.7,  # Min displayed confidence
+    embedding_quality_max_confidence=0.95  # Max displayed confidence
+)
+```
+
+### Handling Irrelevant Queries
+
+The embedding strategy can detect when a query is completely unrelated to the content:
+
+```python
+# This will stop quickly with low confidence
+result = await adaptive.digest(
+    start_url="https://docs.python.org/3/",
+    query="how to cook pasta"  # Irrelevant to Python docs
+)
+
+# Check if query was irrelevant
+if result.metrics.get('is_irrelevant', False):
+    print("Query is unrelated to the content!")
+```
+
+## When to Use Adaptive Crawling
+
+### Perfect For:
+- **Research Tasks**: Finding comprehensive information about a topic
+- **Question Answering**: Gathering sufficient context to answer specific queries
+- **Knowledge Base Building**: Creating focused datasets for AI/ML applications
+- **Competitive Intelligence**: Collecting complete information about specific products/features
+
+### Not Recommended For:
+- **Full Site Archiving**: When you need every page regardless of content
+- **Structured Data Extraction**: When targeting specific, known page patterns
+- **Real-time Monitoring**: When you need continuous updates
+
+## Understanding the Output
+
+### Confidence Score
+
+The confidence score (0-1) indicates how sufficient the gathered information is:
+- **0.0-0.3**: Insufficient information, needs more crawling
+- **0.3-0.6**: Partial information, may answer basic queries
+- **0.6-0.7**: Good coverage, can answer most queries
+- **0.7-1.0**: Excellent coverage, comprehensive information
+
+### Statistics Display
+
+```python
+adaptive.print_stats(detailed=False)  # Summary table
+adaptive.print_stats(detailed=True)   # Detailed metrics
+```
+
+The summary shows:
+- Pages crawled vs. confidence achieved
+- Coverage, consistency, and saturation scores
+- Crawling efficiency metrics
+
+## Persistence and Resumption
+
+### Saving Progress
+
+```python
+config = AdaptiveConfig(
+    save_state=True,
+    state_path="my_crawl_state.json"
+)
+
+# Crawl will auto-save progress
+result = await adaptive.digest(start_url, query)
+```
+
+### Resuming a Crawl
+
+```python
+# Resume from saved state
+result = await adaptive.digest(
+    start_url,
+    query,
+    resume_from="my_crawl_state.json"
+)
+```
+
+### Exporting Knowledge Base
+
+```python
+# Export collected pages to JSONL
+adaptive.export_knowledge_base("knowledge_base.jsonl")
+
+# Import into another session
+new_adaptive = AdaptiveCrawler(crawler)
+new_adaptive.import_knowledge_base("knowledge_base.jsonl")
+```
+
+## Best Practices
+
+### 1. Query Formulation
+- Use specific, descriptive queries
+- Include key terms you expect to find
+- Avoid overly broad queries
+
+### 2. Threshold Tuning
+- Start with default (0.7) for general use
+- Lower to 0.5-0.6 for exploratory crawling
+- Raise to 0.8+ for exhaustive coverage
+
+### 3. Performance Optimization
+- Use appropriate `max_pages` limits
+- Adjust `top_k_links` based on site structure
+- Enable caching for repeat crawls
+
+### 4. Link Selection
+- The crawler prioritizes links based on:
+  - Relevance to query
+  - Expected information gain
+  - URL structure and depth
+
+## Examples
+
+### Research Assistant
+
+```python
+# Gather information about a programming concept
+result = await adaptive.digest(
+    start_url="https://realpython.com",
+    query="python decorators implementation patterns"
+)
+
+# Get the most relevant excerpts
+for doc in adaptive.get_relevant_content(top_k=3):
+    print(f"\nFrom: {doc['url']}")
+    print(f"Relevance: {doc['score']:.2%}")
+    print(doc['content'][:500] + "...")
+```
+
+### Knowledge Base Builder
+
+```python
+# Build a focused knowledge base about machine learning
+queries = [
+    "supervised learning algorithms",
+    "neural network architectures", 
+    "model evaluation metrics"
+]
+
+for query in queries:
+    await adaptive.digest(
+        start_url="https://scikit-learn.org/stable/",
+        query=query
+    )
+    
+# Export combined knowledge base
+adaptive.export_knowledge_base("ml_knowledge.jsonl")
+```
+
+### API Documentation Crawler
+
+```python
+# Intelligently crawl API documentation
+config = AdaptiveConfig(
+    confidence_threshold=0.85,  # Higher threshold for completeness
+    max_pages=30
+)
+
+adaptive = AdaptiveCrawler(crawler, config)
+result = await adaptive.digest(
+    start_url="https://api.example.com/docs",
+    query="authentication endpoints rate limits"
+)
+```
+
+## Next Steps
+
+- Learn about [Advanced Adaptive Strategies](../advanced/adaptive-strategies.md)
+- Explore the [AdaptiveCrawler API Reference](../api/adaptive-crawler.md)
+- See more [Examples](https://github.com/unclecode/crawl4ai/tree/main/docs/examples/adaptive_crawling)
+
+## FAQ
+
+**Q: How is this different from traditional crawling?**
+A: Traditional crawling follows fixed patterns (BFS/DFS). Adaptive crawling makes intelligent decisions about which links to follow and when to stop based on information gain.
+
+**Q: Can I use this with JavaScript-heavy sites?**
+A: Yes! AdaptiveCrawler inherits all capabilities from AsyncWebCrawler, including JavaScript execution.
+
+**Q: How does it handle large websites?**
+A: The algorithm naturally limits crawling to relevant sections. Use `max_pages` as a safety limit.
+
+**Q: Can I customize the scoring algorithms?**
+A: Advanced users can implement custom strategies. See [Adaptive Strategies](../advanced/adaptive-strategies.md).

docs/md_v2/core/browser-crawler-config.md

@@ -29,6 +29,7 @@ class BrowserConfig:
         text_mode=False,
         light_mode=False,
         extra_args=None,
+        enable_stealth=False,
         # ... other advanced parameters omitted here
     ):
         ...
@@ -84,6 +85,11 @@ class BrowserConfig:
     - Additional flags for the underlying browser.  
     - E.g. `["--disable-extensions"]`.
 
+11. **`enable_stealth`**:  
+    - If `True`, enables stealth mode using playwright-stealth.  
+    - Modifies browser fingerprints to avoid basic bot detection.  
+    - Default is `False`. Recommended for sites with bot protection.
+
 ### Helper Methods
 
 Both configuration classes provide a `clone()` method to create modified copies:
@@ -209,7 +215,13 @@ class CrawlerRunConfig:
     - The maximum number of concurrent crawl sessions.  
     - Helps prevent overwhelming the system.
 
-14. **`display_mode`**:  
+14. **`url_matcher`** & **`match_mode`**:  
+    - Enable URL-specific configurations when used with `arun_many()`.
+    - Set `url_matcher` to patterns (glob, function, or list) to match specific URLs.
+    - Use `match_mode` (OR/AND) to control how multiple patterns combine.
+    - See [URL-Specific Configurations](../api/arun_many.md#url-specific-configurations) for examples.
+
+15. **`display_mode`**:  
     - The display mode for progress information (`DETAILED`, `BRIEF`, etc.).  
     - Affects how much information is printed during the crawl.
 
@@ -252,7 +264,7 @@ The `clone()` method:
 ### Key fields to note
 
 1. **`provider`**:  
-- Which LLM provoder to use. 
+- Which LLM provider to use. 
 - Possible values are `"ollama/llama3","groq/llama3-70b-8192","groq/llama3-8b-8192", "openai/gpt-4o-mini" ,"openai/gpt-4o","openai/o1-mini","openai/o1-preview","openai/o3-mini","openai/o3-mini-high","anthropic/claude-3-haiku-20240307","anthropic/claude-3-opus-20240229","anthropic/claude-3-sonnet-20240229","anthropic/claude-3-5-sonnet-20240620","gemini/gemini-pro","gemini/gemini-1.5-pro","gemini/gemini-2.0-flash","gemini/gemini-2.0-flash-exp","gemini/gemini-2.0-flash-lite-preview-02-05","deepseek/deepseek-chat"`<br/>*(default: `"openai/gpt-4o-mini"`)*
 
 2. **`api_token`**:  
@@ -273,7 +285,7 @@ In a typical scenario, you define **one** `BrowserConfig` for your crawler sessi
 
 ```python
 import asyncio
-from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig, CacheMode, LLMConfig
+from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig, CacheMode, LLMConfig, LLMContentFilter, DefaultMarkdownGenerator
 from crawl4ai import JsonCssExtractionStrategy
 
 async def main():
@@ -298,7 +310,7 @@ async def main():
     # 3) Example LLM content filtering
 
     gemini_config = LLMConfig(
-        provider="gemini/gemini-1.5-pro" 
+        provider="gemini/gemini-1.5-pro", 
         api_token = "env:GEMINI_API_TOKEN"
     )
 
@@ -322,8 +334,9 @@ async def main():
     )
 
     md_generator = DefaultMarkdownGenerator(
-    content_filter=filter,
-    options={"ignore_links": True}
+        content_filter=filter,
+        options={"ignore_links": True}
+    )
 
     # 4) Crawler run config: skip cache, use extraction
     run_conf = CrawlerRunConfig(

docs/md_v2/core/c4a-script.md

@@ -52,11 +52,9 @@ That's it! In just a few lines, you've automated a complete search workflow.
 
 Want to learn by doing? We've got you covered:
 
-**🚀 [Live Demo](https://docs.crawl4ai.com/c4a-script/demo)** - Try C4A-Script in your browser right now!
+**🚀 [Live Demo](https://docs.crawl4ai.com/apps/c4a-script/)** - Try C4A-Script in your browser right now!
 
-**📁 [Tutorial Examples](/examples/c4a_script/)** - Complete examples with source code
-
-**🛠️ [Local Tutorial](/examples/c4a_script/tutorial/)** - Run the interactive tutorial on your machine
+**📁 [Tutorial Examples](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/c4a_script/)** - Complete examples with source code
 
 ### Running the Tutorial Locally
 

docs/md_v2/core/cli.md

@@ -17,6 +17,9 @@
 - [Configuration Reference](#configuration-reference)
 - [Best Practices & Tips](#best-practices--tips)
 
+## Installation
+The Crawl4AI CLI will be installed automatically when you install the library.
+
 ## Basic Usage
 
 The Crawl4AI CLI (`crwl`) provides a simple interface to the Crawl4AI library:

docs/md_v2/core/content-selection.md

@@ -350,15 +350,22 @@ if __name__ == "__main__":
 
 ## 6. Scraping Modes
 
-Crawl4AI provides two different scraping strategies for HTML content processing: `WebScrapingStrategy` (BeautifulSoup-based, default) and `LXMLWebScrapingStrategy` (LXML-based). The LXML strategy offers significantly better performance, especially for large HTML documents.
+Crawl4AI uses `LXMLWebScrapingStrategy` (LXML-based) as the default scraping strategy for HTML content processing. This strategy offers excellent performance, especially for large HTML documents.
+
+**Note:** For backward compatibility, `WebScrapingStrategy` is still available as an alias for `LXMLWebScrapingStrategy`.
 
 ```python
 from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, LXMLWebScrapingStrategy
 
 async def main():
-    config = CrawlerRunConfig(
-        scraping_strategy=LXMLWebScrapingStrategy()  # Faster alternative to default BeautifulSoup
+    # Default configuration already uses LXMLWebScrapingStrategy
+    config = CrawlerRunConfig()
+    
+    # Or explicitly specify it if desired
+    config_explicit = CrawlerRunConfig(
+        scraping_strategy=LXMLWebScrapingStrategy()
     )
+    
     async with AsyncWebCrawler() as crawler:
         result = await crawler.arun(
             url="https://example.com", 
@@ -417,21 +424,20 @@ class CustomScrapingStrategy(ContentScrapingStrategy):
 
 ### Performance Considerations
 
-The LXML strategy can be up to 10-20x faster than BeautifulSoup strategy, particularly when processing large HTML documents. However, please note:
+The LXML strategy provides excellent performance, particularly when processing large HTML documents, offering up to 10-20x faster processing compared to BeautifulSoup-based approaches.
 
-1. LXML strategy is currently experimental
-2. In some edge cases, the parsing results might differ slightly from BeautifulSoup
-3. If you encounter any inconsistencies between LXML and BeautifulSoup results, please [raise an issue](https://github.com/codeium/crawl4ai/issues) with a reproducible example
+Benefits of LXML strategy:
+- Fast processing of large HTML documents (especially >100KB)
+- Efficient memory usage
+- Good handling of well-formed HTML
+- Robust table detection and extraction
 
-Choose LXML strategy when:
-- Processing large HTML documents (recommended for >100KB)
-- Performance is critical
-- Working with well-formed HTML
+### Backward Compatibility
 
-Stick to BeautifulSoup strategy (default) when:
-- Maximum compatibility is needed
-- Working with malformed HTML
-- Exact parsing behavior is critical
+For users upgrading from earlier versions:
+- `WebScrapingStrategy` is now an alias for `LXMLWebScrapingStrategy`
+- Existing code using `WebScrapingStrategy` will continue to work without modification
+- No changes are required to your existing code
 
 ---
 

docs/md_v2/core/crawler-result.md

@@ -19,13 +19,15 @@ class MarkdownGenerationResult(BaseModel):
 class CrawlResult(BaseModel):
     url: str
     html: str
+    fit_html: Optional[str] = None
     success: bool
     cleaned_html: Optional[str] = None
     media: Dict[str, List[Dict]] = {}
     links: Dict[str, List[Dict]] = {}
     downloaded_files: Optional[List[str]] = None
+    js_execution_result: Optional[Dict[str, Any]] = None
     screenshot: Optional[str] = None
-    pdf : Optional[bytes] = None
+    pdf: Optional[bytes] = None
     mhtml: Optional[str] = None
     markdown: Optional[Union[str, MarkdownGenerationResult]] = None
     extracted_content: Optional[str] = None
@@ -35,6 +37,12 @@ class CrawlResult(BaseModel):
     response_headers: Optional[dict] = None
     status_code: Optional[int] = None
     ssl_certificate: Optional[SSLCertificate] = None
+    dispatch_result: Optional[DispatchResult] = None
+    redirected_url: Optional[str] = None
+    network_requests: Optional[List[Dict[str, Any]]] = None
+    console_messages: Optional[List[Dict[str, Any]]] = None
+    tables: List[Dict] = Field(default_factory=list)
+
     class Config:
         arbitrary_types_allowed = True
 ```
@@ -45,11 +53,13 @@ class CrawlResult(BaseModel):
 |-------------------------------------------|-----------------------------------------------------------------------------------------------------|
 | **url (`str`)**                           | The final or actual URL crawled (in case of redirects).                                             |
 | **html (`str`)**                          | Original, unmodified page HTML. Good for debugging or custom processing.                            |
+| **fit_html (`Optional[str]`)**            | Preprocessed HTML optimized for extraction and content filtering.                                    |
 | **success (`bool`)**                      | `True` if the crawl completed without major errors, else `False`.                                   |
 | **cleaned_html (`Optional[str]`)**        | Sanitized HTML with scripts/styles removed; can exclude tags if configured via `excluded_tags` etc. |
 | **media (`Dict[str, List[Dict]]`)**       | Extracted media info (images, audio, etc.), each with attributes like `src`, `alt`, `score`, etc.   |
 | **links (`Dict[str, List[Dict]]`)**       | Extracted link data, split by `internal` and `external`. Each link usually has `href`, `text`, etc. |
 | **downloaded_files (`Optional[List[str]]`)** | If `accept_downloads=True` in `BrowserConfig`, this lists the filepaths of saved downloads.         |
+| **js_execution_result (`Optional[Dict[str, Any]]`)** | Results from JavaScript execution during crawling. |
 | **screenshot (`Optional[str]`)**          | Screenshot of the page (base64-encoded) if `screenshot=True`.                                       |
 | **pdf (`Optional[bytes]`)**               | PDF of the page if `pdf=True`.                                                                      |
 | **mhtml (`Optional[str]`)**               | MHTML snapshot of the page if `capture_mhtml=True`. Contains the full page with all resources.      |
@@ -61,6 +71,11 @@ class CrawlResult(BaseModel):
 | **response_headers (`Optional[dict]`)**   | HTTP response headers, if captured.                                                                 |
 | **status_code (`Optional[int]`)**         | HTTP status code (e.g., 200 for OK).                                                                |
 | **ssl_certificate (`Optional[SSLCertificate]`)** | SSL certificate info if `fetch_ssl_certificate=True`.                                               |
+| **dispatch_result (`Optional[DispatchResult]`)** | Additional concurrency and resource usage information when crawling URLs in parallel.               |
+| **redirected_url (`Optional[str]`)**      | The URL after any redirects (different from `url` which is the final URL).                          |
+| **network_requests (`Optional[List[Dict[str, Any]]]`)** | List of network requests, responses, and failures captured during the crawl if `capture_network_requests=True`. |
+| **console_messages (`Optional[List[Dict[str, Any]]]`)** | List of browser console messages captured during the crawl if `capture_console_messages=True`.       |
+| **tables (`List[Dict]`)**                 | Table data extracted from HTML tables with structure `[{headers, rows, caption, summary}]`.           |
 
 ---
 
@@ -172,7 +187,7 @@ Here:
 
 ---
 
-## 5. More Fields: Links, Media, and More
+## 5. More Fields: Links, Media, Tables and More
 
 ### 5.1 `links`
 
@@ -192,7 +207,77 @@ for img in images:
     print("Image URL:", img["src"], "Alt:", img.get("alt"))
 ```
 
-### 5.3 `screenshot`, `pdf`, and `mhtml`
+### 5.3 `tables`
+
+The `tables` field contains structured data extracted from HTML tables found on the crawled page. Tables are analyzed based on various criteria to determine if they are actual data tables (as opposed to layout tables), including:
+
+- Presence of thead and tbody sections
+- Use of th elements for headers
+- Column consistency
+- Text density
+- And other factors
+
+Tables that score above the threshold (default: 7) are extracted and stored in result.tables.
+
+### Accessing Table data:
+```python
+import asyncio
+from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
+
+async def main():
+    async with AsyncWebCrawler() as crawler:
+        result = await crawler.arun(
+            url="https://www.w3schools.com/html/html_tables.asp",
+            config=CrawlerRunConfig(
+                table_score_threshold=7  # Minimum score for table detection
+            )
+        )
+
+        if result.success and result.tables:
+            print(f"Found {len(result.tables)} tables")
+
+            for i, table in enumerate(result.tables):
+                print(f"\nTable {i+1}:")
+                print(f"Caption: {table.get('caption', 'No caption')}")
+                print(f"Headers: {table['headers']}")
+                print(f"Rows: {len(table['rows'])}")
+
+                # Print first few rows as example
+                for j, row in enumerate(table['rows'][:3]):
+                    print(f"  Row {j+1}: {row}")
+
+if __name__ == "__main__":
+    asyncio.run(main())
+
+```
+
+### Configuring Table Extraction:
+
+You can adjust the sensitivity of the table detection algorithm with:
+
+```python
+config = CrawlerRunConfig(
+    table_score_threshold=5  # Lower value = more tables detected (default: 7)
+)
+```
+
+Each extracted table contains: 
+
+- `headers`: Column header names 
+- `rows`: List of rows, each containing cell values
+- `caption`: Table caption text (if available) 
+- `summary`: Table summary attribute (if specified)
+
+### Table Extraction Tips
+
+- Not all HTML tables are extracted - only those detected as "data tables" vs. layout tables.
+- Tables with inconsistent cell counts, nested tables, or those used purely for layout may be skipped.
+- If you're missing tables, try adjusting the `table_score_threshold` to a lower value (default is 7).
+
+The table detection algorithm scores tables based on features like consistent columns, presence of headers, text density, and more. Tables scoring above the threshold are considered data tables worth extracting.
+
+
+### 5.4 `screenshot`, `pdf`, and `mhtml`
 
 If you set `screenshot=True`, `pdf=True`, or `capture_mhtml=True` in **`CrawlerRunConfig`**, then:
 
@@ -213,7 +298,7 @@ if result.mhtml:
 
 The MHTML (MIME HTML) format is particularly useful as it captures the entire web page including all of its resources (CSS, images, scripts, etc.) in a single file, making it perfect for archiving or offline viewing.
 
-### 5.4 `ssl_certificate`
+### 5.5 `ssl_certificate`
 
 If `fetch_ssl_certificate=True`, `result.ssl_certificate` holds details about the site’s SSL cert, such as issuer, validity dates, etc.
 

docs/md_v2/core/docker-deployment.md

@@ -58,13 +58,15 @@ Pull and run images directly from Docker Hub without building locally.
 
 #### 1. Pull the Image
 
-Our latest release candidate is `0.6.0-r2`. Images are built with multi-arch manifests, so Docker automatically pulls the correct version for your system.
+Our latest release candidate is `0.7.0-r1`. Images are built with multi-arch manifests, so Docker automatically pulls the correct version for your system.
+
+> ⚠️ **Important Note**: The `latest` tag currently points to the stable `0.6.0` version. After testing and validation, `0.7.0` (without -r1) will be released and `latest` will be updated. For now, please use `0.7.0-r1` to test the new features.
 
 ```bash
-# Pull the release candidate (recommended for latest features)
-docker pull unclecode/crawl4ai:0.6.0-r1
+# Pull the release candidate (for testing new features)
+docker pull unclecode/crawl4ai:0.7.0-r1
 
-# Or pull the latest stable version
+# Or pull the current stable version (0.6.0)
 docker pull unclecode/crawl4ai:latest
 ```
 
@@ -124,7 +126,7 @@ docker stop crawl4ai && docker rm crawl4ai
 #### Docker Hub Versioning Explained
 
 *   **Image Name:** `unclecode/crawl4ai`
-*   **Tag Format:** `LIBRARY_VERSION[-SUFFIX]` (e.g., `0.6.0-r2`)
+*   **Tag Format:** `LIBRARY_VERSION[-SUFFIX]` (e.g., `0.7.0-r1`)
     *   `LIBRARY_VERSION`: The semantic version of the core `crawl4ai` Python library
     *   `SUFFIX`: Optional tag for release candidates (``) and revisions (`r1`)
 *   **`latest` Tag:** Points to the most recent stable version
@@ -152,6 +154,30 @@ cp deploy/docker/.llm.env.example .llm.env
 # Now edit .llm.env and add your API keys
 ```
 
+**Flexible LLM Provider Configuration:**
+
+The Docker setup now supports flexible LLM provider configuration through three methods:
+
+1. **Environment Variable** (Highest Priority): Set `LLM_PROVIDER` to override the default
+   ```bash
+   export LLM_PROVIDER="anthropic/claude-3-opus"
+   # Or in your .llm.env file:
+   # LLM_PROVIDER=anthropic/claude-3-opus
+   ```
+
+2. **API Request Parameter**: Specify provider per request
+   ```json
+   {
+     "url": "https://example.com",
+     "f": "llm",
+     "provider": "groq/mixtral-8x7b"
+   }
+   ```
+
+3. **Config File Default**: Falls back to `config.yml` (default: `openai/gpt-4o-mini`)
+
+The system automatically selects the appropriate API key based on the configured `api_key_env` in the config file.
+
 #### 3. Build and Run with Compose
 
 The `docker-compose.yml` file in the project root provides a simplified approach that automatically handles architecture detection using buildx.
@@ -666,7 +692,7 @@ app:
 
 # Default LLM Configuration
 llm:
-  provider: "openai/gpt-4o-mini"
+  provider: "openai/gpt-4o-mini"  # Can be overridden by LLM_PROVIDER env var
   api_key_env: "OPENAI_API_KEY"
   # api_key: sk-...  # If you pass the API key directly then api_key_env will be ignored
 

docs/md_v2/core/examples.md

@@ -29,6 +29,7 @@ This page provides a comprehensive list of example scripts that demonstrate vari
 |---------|-------------|------|
 | Deep Crawling | An extensive tutorial on deep crawling capabilities, demonstrating BFS and BestFirst strategies, stream vs. non-stream execution, filters, scorers, and advanced configurations. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/deepcrawl_example.py) |
 | Virtual Scroll | Comprehensive examples for handling virtualized scrolling on sites like Twitter, Instagram. Demonstrates different scrolling scenarios with local test server. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/virtual_scroll_example.py) |
+| Adaptive Crawling | Demonstrates intelligent crawling that automatically determines when sufficient information has been gathered. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/adaptive_crawling/) |
 | Dispatcher | Shows how to use the crawl dispatcher for advanced workload management. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/dispatcher_example.py) |
 | Storage State | Tutorial on managing browser storage state for persistence. | [View Guide](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/storage_state_tutorial.md) |
 | Network Console Capture | Demonstrates how to capture and analyze network requests and console logs. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/network_console_capture_example.py) |
@@ -53,6 +54,16 @@ This page provides a comprehensive list of example scripts that demonstrate vari
 | Crypto Analysis | Demonstrates how to crawl and analyze cryptocurrency data. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/crypto_analysis_example.py) |
 | SERP API | Demonstrates using Crawl4AI with search engine result pages. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/serp_api_project_11_feb.py) |
 
+## Anti-Bot & Stealth Features
+
+| Example | Description | Link |
+|---------|-------------|------|
+| Stealth Mode Quick Start | Five practical examples showing how to use stealth mode for bypassing basic bot detection. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/stealth_mode_quick_start.py) |
+| Stealth Mode Comprehensive | Comprehensive demonstration of stealth mode features with bot detection testing and comparisons. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/stealth_mode_example.py) |
+| Undetected Browser | Simple example showing how to use the undetected browser adapter. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/hello_world_undetected.py) |
+| Undetected Browser Demo | Basic demo comparing regular and undetected browser modes. | [View Code](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/undetected_simple_demo.py) |
+| Undetected Tests | Advanced tests comparing regular vs undetected browsers on various bot detection services. | [View Folder](https://github.com/unclecode/crawl4ai/tree/main/docs/examples/undetectability/) |
+
 ## Customization & Security
 
 | Example | Description | Link |
@@ -113,4 +124,4 @@ Some examples may require:
 
 ## Contributing New Examples
 
-If you've created an interesting example that demonstrates a unique use case or feature of Crawl4AI, we encourage you to contribute it to our examples collection. Please see our [contribution guidelines](https://github.com/unclecode/crawl4ai/blob/main/CONTRIBUTORS.md) for more information.
+If you've created an interesting example that demonstrates a unique use case or feature of Crawl4AI, we encourage you to contribute it to our examples collection. Please see our [contribution guidelines](https://github.com/unclecode/crawl4ai/blob/main/CONTRIBUTORS.md) for more information.

docs/md_v2/core/installation.md

@@ -18,7 +18,7 @@ crawl4ai-setup
 ```
 
 **What does it do?**
-- Installs or updates required Playwright browsers (Chromium, Firefox, etc.)
+- Installs or updates required browser dependencies for both regular and undetected modes
 - Performs OS-level checks (e.g., missing libs on Linux)
 - Confirms your environment is ready to crawl
 

docs/md_v2/core/link-media.md

@@ -125,7 +125,7 @@ Here's a full example you can copy, paste, and run immediately:
 ```python
 import asyncio
 from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
-from crawl4ai.async_configs import LinkPreviewConfig
+from crawl4ai import LinkPreviewConfig
 
 async def extract_link_heads_example():
     """
@@ -237,7 +237,7 @@ if __name__ == "__main__":
 The `LinkPreviewConfig` class supports these options:
 
 ```python
-from crawl4ai.async_configs import LinkPreviewConfig
+from crawl4ai import LinkPreviewConfig
 
 link_preview_config = LinkPreviewConfig(
     # BASIC SETTINGS
@@ -520,7 +520,8 @@ This approach is handy when you still want external links but need to block cert
 
 ### 4.1 Accessing `result.media`
 
-By default, Crawl4AI collects images, audio, video URLs, and data tables it finds on the page. These are stored in `result.media`, a dictionary keyed by media type (e.g., `images`, `videos`, `audio`, `tables`).
+By default, Crawl4AI collects images, audio and video URLs it finds on the page. These are stored in `result.media`, a dictionary keyed by media type (e.g., `images`, `videos`, `audio`).
+**Note: Tables have been moved from `result.media["tables"]` to the new `result.tables` format for better organization and direct access.**
 
 **Basic Example**:
 
@@ -534,14 +535,6 @@ if result.success:
         print(f"           Alt text: {img.get('alt', '')}")
         print(f"           Score: {img.get('score')}")
         print(f"           Description: {img.get('desc', '')}\n")
-    
-    # Get tables
-    tables = result.media.get("tables", [])
-    print(f"Found {len(tables)} data tables in total.")
-    for i, table in enumerate(tables):
-        print(f"[Table {i}] Caption: {table.get('caption', 'No caption')}")
-        print(f"           Columns: {len(table.get('headers', []))}")
-        print(f"           Rows: {len(table.get('rows', []))}")
 ```
 
 **Structure Example**:
@@ -568,19 +561,6 @@ result.media = {
   "audio": [
     # Similar structure but with audio-specific fields
   ],
-  "tables": [
-    {
-      "headers": ["Name", "Age", "Location"],
-      "rows": [
-        ["John Doe", "34", "New York"],
-        ["Jane Smith", "28", "San Francisco"],
-        ["Alex Johnson", "42", "Chicago"]
-      ],
-      "caption": "Employee Directory",
-      "summary": "Directory of company employees"
-    },
-    # More tables if present
-  ]
 }
 ```
 
@@ -608,53 +588,7 @@ crawler_cfg = CrawlerRunConfig(
 
 This setting attempts to discard images from outside the primary domain, keeping only those from the site you’re crawling.
 
-### 3.3 Working with Tables
-
-Crawl4AI can detect and extract structured data from HTML tables. Tables are analyzed based on various criteria to determine if they are actual data tables (as opposed to layout tables), including:
-
-- Presence of thead and tbody sections
-- Use of th elements for headers
-- Column consistency
-- Text density
-- And other factors
-
-Tables that score above the threshold (default: 7) are extracted and stored in `result.media.tables`.
-
-**Accessing Table Data**:
-
-```python
-if result.success:
-    tables = result.media.get("tables", [])
-    print(f"Found {len(tables)} data tables on the page")
-    
-    if tables:
-        # Access the first table
-        first_table = tables[0]
-        print(f"Table caption: {first_table.get('caption', 'No caption')}")
-        print(f"Headers: {first_table.get('headers', [])}")
-        
-        # Print the first 3 rows
-        for i, row in enumerate(first_table.get('rows', [])[:3]):
-            print(f"Row {i+1}: {row}")
-```
-
-**Configuring Table Extraction**:
-
-You can adjust the sensitivity of the table detection algorithm with:
-
-```python
-crawler_cfg = CrawlerRunConfig(
-    table_score_threshold=5  # Lower value = more tables detected (default: 7)
-)
-```
-
-Each extracted table contains:
-- `headers`: Column header names
-- `rows`: List of rows, each containing cell values
-- `caption`: Table caption text (if available)
-- `summary`: Table summary attribute (if specified)
-
-### 3.4 Additional Media Config
+### 4.3 Additional Media Config
 
 - **`screenshot`**: Set to `True` if you want a full-page screenshot stored as `base64` in `result.screenshot`.  
 - **`pdf`**: Set to `True` if you want a PDF version of the page in `result.pdf`.  
@@ -695,7 +629,7 @@ The MHTML format is particularly useful because:
 
 ---
 
-## 4. Putting It All Together: Link & Media Filtering
+## 5. Putting It All Together: Link & Media Filtering
 
 Here’s a combined example demonstrating how to filter out external links, skip certain domains, and exclude external images:
 
@@ -743,7 +677,7 @@ if __name__ == "__main__":
 
 ---
 
-## 5. Common Pitfalls & Tips
+## 6. Common Pitfalls & Tips
 
 1. **Conflicting Flags**:  
    - `exclude_external_links=True` but then also specifying `exclude_social_media_links=True` is typically fine, but understand that the first setting already discards *all* external links. The second becomes somewhat redundant.  
@@ -762,10 +696,3 @@ if __name__ == "__main__":
 ---
 
 **That’s it for Link & Media Analysis!** You’re now equipped to filter out unwanted sites and zero in on the images and videos that matter for your project.
-### Table Extraction Tips
-
-- Not all HTML tables are extracted - only those detected as "data tables" vs. layout tables.
-- Tables with inconsistent cell counts, nested tables, or those used purely for layout may be skipped.
-- If you're missing tables, try adjusting the `table_score_threshold` to a lower value (default is 7).
-
-The table detection algorithm scores tables based on features like consistent columns, presence of headers, text density, and more. Tables scoring above the threshold are considered data tables worth extracting.