feat: Enhance stealth compatibility with new and legacy APIs, add configuration support

Fixes #1273

- Replace deprecated stealth_async import with Stealth class
- Add stealth flag to BrowserConfig (default: true)
- Update async_crawler_strategy to use Stealth().apply_stealth_async()
- Remove obsolete StealthConfig from browser_manager
- Maintain backward compatibility with existing stealth functionality

This fixes compatibility issues with playwright-stealth v2.0.0+ where the API changed from stealth_async function to Stealth class.

test: Add comprehensive tests for playwright-stealth v2.0.0+ compatibility

- Test Stealth class import and instantiation
- Test apply_stealth_async method availability
- Test BrowserConfig stealth flag functionality
- Test stealth flag serialization
- Verify backward compatibility with existing stealth functionality

README.md

@@ -523,18 +523,15 @@ async def test_news_crawl():
 - **🧠 Adaptive Crawling**: Your crawler now learns and adapts to website patterns automatically:
   ```python
   config = AdaptiveConfig(
-      confidence_threshold=0.7, # Min confidence to stop crawling
+      confidence_threshold=0.7,
-      max_depth=5, # Maximum crawl depth
+      max_history=100,
-      max_pages=20, # Maximum number of pages to crawl
+      learning_rate=0.2
-      strategy="statistical"
   )
   
-  async with AsyncWebCrawler() as crawler:
+  result = await crawler.arun(
-      adaptive_crawler = AdaptiveCrawler(crawler, config)
+      "https://news.example.com",
-      state = await adaptive_crawler.digest(
+      config=CrawlerRunConfig(adaptive_config=config)
-          start_url="https://news.example.com",
+  )
-          query="latest news content"
-      )
   # Crawler learns patterns and improves extraction over time
   ```
 

crawl4ai/async_crawler_strategy.py

@@ -12,6 +12,20 @@ from playwright.async_api import TimeoutError as PlaywrightTimeoutError
 from io import BytesIO
 from PIL import Image, ImageDraw, ImageFont
 import hashlib
+
+# Backward compatible stealth import
+try:
+    # Try new tf-playwright-stealth API (Stealth class)
+    from playwright_stealth import Stealth
+    STEALTH_NEW_API = True
+except ImportError:
+    try:
+        # Try old playwright-stealth API (stealth_async function)
+        from playwright_stealth import stealth_async
+        STEALTH_NEW_API = False
+    except ImportError:
+        # No stealth available
+        STEALTH_NEW_API = None
 import uuid
 from .js_snippet import load_js_script
 from .models import AsyncCrawlResponse
@@ -31,6 +45,107 @@ from types import MappingProxyType
 import contextlib
 from functools import partial
 
+
+# Add StealthConfig class for backward compatibility and new features
+class StealthConfig:
+    """
+    Configuration class for stealth settings that works with tf-playwright-stealth.
+    This maintains backward compatibility while supporting all tf-playwright-stealth features.
+    """
+    def __init__(
+        self,
+        # Common settings
+        enabled: bool = True,
+        
+        # Core tf-playwright-stealth parameters (matching the actual library)
+        chrome_app: bool = True,
+        chrome_csi: bool = True,
+        chrome_load_times: bool = True,
+        chrome_runtime: bool = False,  # Note: library default is False
+        hairline: bool = True,
+        iframe_content_window: bool = True,
+        media_codecs: bool = True,
+        navigator_hardware_concurrency: bool = True,
+        navigator_languages: bool = True,
+        navigator_permissions: bool = True,
+        navigator_platform: bool = True,
+        navigator_plugins: bool = True,
+        navigator_user_agent: bool = True,
+        navigator_vendor: bool = True,
+        navigator_webdriver: bool = True,
+        sec_ch_ua: bool = True,
+        webgl_vendor: bool = True,
+        
+        # Override parameters
+        navigator_languages_override: tuple = ("en-US", "en"),
+        navigator_platform_override: str = "Win32",
+        navigator_user_agent_override: str = None,
+        navigator_vendor_override: str = None,
+        sec_ch_ua_override: str = None,
+        webgl_renderer_override: str = None,
+        webgl_vendor_override: str = None,
+        
+        # Advanced parameters
+        init_scripts_only: bool = False,
+        script_logging: bool = False,
+        
+        # Legacy parameters for backward compatibility
+        webdriver: bool = None,  # This will be mapped to navigator_webdriver
+        user_agent_override: bool = None,  # This will be mapped to navigator_user_agent
+        window_outerdimensions: bool = None,  # This parameter doesn't exist in tf-playwright-stealth
+    ):
+        self.enabled = enabled
+        
+        # Handle legacy parameter mapping for backward compatibility
+        if webdriver is not None:
+            navigator_webdriver = webdriver
+        if user_agent_override is not None:
+            navigator_user_agent = user_agent_override
+        
+        # Store all stealth options for the Stealth class - filter out None values
+        self.stealth_options = {
+            k: v for k, v in {
+                'chrome_app': chrome_app,
+                'chrome_csi': chrome_csi,
+                'chrome_load_times': chrome_load_times,
+                'chrome_runtime': chrome_runtime,
+                'hairline': hairline,
+                'iframe_content_window': iframe_content_window,
+                'media_codecs': media_codecs,
+                'navigator_hardware_concurrency': navigator_hardware_concurrency,
+                'navigator_languages': navigator_languages,
+                'navigator_permissions': navigator_permissions,
+                'navigator_platform': navigator_platform,
+                'navigator_plugins': navigator_plugins,
+                'navigator_user_agent': navigator_user_agent,
+                'navigator_vendor': navigator_vendor,
+                'navigator_webdriver': navigator_webdriver,
+                'sec_ch_ua': sec_ch_ua,
+                'webgl_vendor': webgl_vendor,
+                'navigator_languages_override': navigator_languages_override,
+                'navigator_platform_override': navigator_platform_override,
+                'navigator_user_agent_override': navigator_user_agent_override,
+                'navigator_vendor_override': navigator_vendor_override,
+                'sec_ch_ua_override': sec_ch_ua_override,
+                'webgl_renderer_override': webgl_renderer_override,
+                'webgl_vendor_override': webgl_vendor_override,
+                'init_scripts_only': init_scripts_only,
+                'script_logging': script_logging,
+            }.items() if v is not None
+        }
+    
+    @classmethod
+    def from_dict(cls, config_dict: dict) -> 'StealthConfig':
+        """Create StealthConfig from dictionary for easy configuration"""
+        return cls(**config_dict)
+    
+    def to_dict(self) -> dict:
+        """Convert to dictionary for serialization"""
+        return {
+            'enabled': self.enabled,
+            **self.stealth_options
+        }
+
 class AsyncCrawlerStrategy(ABC):
     """
     Abstract base class for crawler strategies.
@@ -39,7 +154,7 @@ class AsyncCrawlerStrategy(ABC):
 
     @abstractmethod
     async def crawl(self, url: str, **kwargs) -> AsyncCrawlResponse:
-        pass  # 4 + 3
+        pass # 4 + 3
 
 class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
     """
@@ -220,6 +335,79 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         """
         self.headers = headers
 
+    async def _apply_stealth(self, page: Page, stealth_config: Optional[StealthConfig] = None):
+        """
+        Apply stealth measures to the page with backward compatibility and enhanced configuration.
+        
+        This method automatically applies stealth measures and now supports configuration
+        through StealthConfig while maintaining backward compatibility.
+        
+        Currently supports:
+        - tf-playwright-stealth (Stealth class with extensive configuration)
+        - Old playwright-stealth v1.x (stealth_async function) - legacy support
+        
+        Args:
+            page (Page): The Playwright page object
+            stealth_config (Optional[StealthConfig]): Configuration for stealth settings
+        """
+        if STEALTH_NEW_API is None:
+            # No stealth library available - silently continue
+            if self.logger and hasattr(self.logger, 'debug'):
+                self.logger.debug(
+                    message="playwright-stealth not available, skipping stealth measures",
+                    tag="STEALTH"
+                )
+            return
+        
+        # Use default config if none provided
+        if stealth_config is None:
+            stealth_config = StealthConfig()
+        
+        # Skip if stealth is disabled
+        if not stealth_config.enabled:
+            if self.logger and hasattr(self.logger, 'debug'):
+                self.logger.debug(
+                    message="Stealth measures disabled in configuration",
+                    tag="STEALTH"
+                )
+            return
+        
+        try:
+            if STEALTH_NEW_API:
+                # Use tf-playwright-stealth API with configuration support
+                # Filter out any invalid parameters that might cause issues
+                valid_options = {}
+                for key, value in stealth_config.stealth_options.items():
+                    # Accept boolean parameters and specific string/tuple parameters
+                    if isinstance(value, (bool, str, tuple)):
+                        valid_options[key] = value
+                
+                stealth = Stealth(**valid_options)
+                await stealth.apply_stealth_async(page)
+                
+                config_info = f"with {len(valid_options)} options"
+            else:
+                # Use old API (v1.x) - configuration options are limited
+                await stealth_async(page)
+                config_info = "default (v1.x legacy)"
+                
+            # Only log if logger is available and in debug mode
+            if self.logger and hasattr(self.logger, 'debug'):
+                api_version = "tf-playwright-stealth" if STEALTH_NEW_API else "v1.x"
+                self.logger.debug(
+                    message="Applied stealth measures using {version} {config}",
+                    tag="STEALTH",
+                    params={"version": api_version, "config": config_info}
+                )
+        except Exception as e:
+            # Silently continue if stealth fails - don't break the crawling process
+            if self.logger:
+                self.logger.warning(
+                    message="Stealth measures failed, continuing without stealth: {error}",
+                    tag="STEALTH",
+                    params={"error": str(e)}
+                )
+
     async def smart_wait(self, page: Page, wait_for: str, timeout: float = 30000):
         """
         Wait for a condition in a smart way. This functions works as below:
@@ -532,6 +720,24 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         # Get page for session
         page, context = await self.browser_manager.get_page(crawlerRunConfig=config)
 
+        # Apply stealth measures automatically (backward compatible) with optional config
+        # Check multiple possible locations for stealth config for flexibility
+        stealth_config = None
+        if hasattr(config, 'stealth_config') and config.stealth_config:
+            stealth_config = config.stealth_config
+        elif hasattr(config, 'stealth') and config.stealth:
+            # Alternative attribute name for backward compatibility
+            stealth_config = config.stealth if isinstance(config.stealth, StealthConfig) else StealthConfig.from_dict(config.stealth)
+        elif config.magic:
+            # Enable more aggressive stealth in magic mode
+            stealth_config = StealthConfig(
+                navigator_webdriver=False,  # More aggressive stealth
+                webdriver=False,
+                chrome_app=False
+            )
+        
+        await self._apply_stealth(page, stealth_config)
+
         # await page.goto(URL)
 
         # Add default cookie
@@ -933,7 +1139,6 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                         tag="VIEWPORT",
                         params={"error": str(e)},
                     )
-
             # Handle full page scanning
             if config.scan_full_page:
                 # await self._handle_full_page_scan(page, config.scroll_delay)
@@ -1837,8 +2042,6 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                         #     }}
                         # }})();
                         # """
-                        # )
-                        
                         # """ NEW VERSION:
                         # When {script} contains statements (e.g., const link = …; link.click();), 
                         # this forms invalid JavaScript, causing Playwright execution error: SyntaxError: Unexpected token 'const'.

crawl4ai/browser_manager.py

@@ -14,24 +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",
     "--disable-background-timer-throttling",

docs/blog/release-v0.7.0.md

@@ -10,8 +10,9 @@ Today I'm releasing Crawl4AI v0.7.0—the Adaptive Intelligence Update. This rel
 
 - **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
+- **Link Preview with 3-Layer Scoring**: Intelligent link analysis and prioritization
 - **Async URL Seeder**: Discover thousands of URLs in seconds with intelligent filtering
+- **PDF Parsing**: Extract data from PDF documents
 - **Performance Optimizations**: Significant speed and memory improvements
 
 ## 🧠 Adaptive Crawling: Intelligence Through Pattern Learning
@@ -29,34 +30,44 @@ The Adaptive Crawler maintains a persistent state for each domain, tracking:
 - Extraction confidence scores
 
 ```python
-from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+from crawl4ai import AdaptiveCrawler, AdaptiveConfig, CrawlState
 
-# Initialize with custom adaptive parameters
+# Initialize with custom learning parameters
 config = AdaptiveConfig(
-    confidence_threshold=0.7,    # Min confidence to stop crawling
+    confidence_threshold=0.7,    # Min confidence to use learned patterns
-    max_depth=5,                # Maximum crawl depth
+    max_history=100,            # Remember last 100 crawls per domain
-    max_pages=20,               # Maximum number of pages to crawl
+    learning_rate=0.2,          # How quickly to adapt to changes
-    top_k_links=3,              # Number of top links to follow per page
+    patterns_per_page=3,        # Patterns to learn per page type
-    strategy="statistical",     # 'statistical' or 'embedding'
+    extraction_strategy='css'   # 'css' or 'xpath'
-    coverage_weight=0.4,        # Weight for coverage in confidence calculation
-    consistency_weight=0.3,     # Weight for consistency in confidence calculation
-    saturation_weight=0.3       # Weight for saturation in confidence calculation
 )
 
-# Initialize adaptive crawler with web crawler
+adaptive_crawler = AdaptiveCrawler(config)
+
+# First crawl - crawler learns the structure
 async with AsyncWebCrawler() as crawler:
-    adaptive_crawler = AdaptiveCrawler(crawler, config)
+    result = await crawler.arun(
-    
+        "https://news.example.com/article/12345",
-    # Crawl and learn patterns
+        config=CrawlerRunConfig(
-    state = await adaptive_crawler.digest(
+            adaptive_config=config,
-        start_url="https://news.example.com/article/12345",
+            extraction_hints={  # Optional hints to speed up learning
-        query="latest news articles and content"
+                "title": "article h1",
+                "content": "article .body-content"
+            }
+        )
     )
     
-    # Access results and confidence
+    # Crawler identifies and stores patterns
-    print(f"Confidence Level: {adaptive_crawler.confidence:.0%}")
+    if result.success:
-    print(f"Pages Crawled: {len(state.crawled_urls)}")
+        state = adaptive_crawler.get_state("news.example.com")
-    print(f"Knowledge Base: {len(adaptive_crawler.state.knowledge_base)} documents")
+        print(f"Learned {len(state.patterns)} patterns")
+        print(f"Confidence: {state.avg_confidence:.2%}")
+
+# Subsequent crawls - uses learned patterns
+result2 = await crawler.arun(
+    "https://news.example.com/article/67890",
+    config=CrawlerRunConfig(adaptive_config=config)
+)
+# Automatically extracts using learned patterns!
 ```
 
 **Expected Real-World Impact:**
@@ -81,7 +92,9 @@ 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
+    wait_after_scroll=1.0,             # Let content load
+    capture_method="incremental",       # Capture new content on each scroll
+    deduplicate=True                   # Remove duplicate elements
 )
 
 # For e-commerce product grids (Instagram style)
@@ -89,7 +102,8 @@ 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
+    wait_after_scroll=1.5,             # Images need time
+    stop_on_no_change=True            # Smart stopping
 )
 
 # For news feeds with lazy loading
@@ -97,7 +111,9 @@ 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
+    wait_after_scroll=0.5,
+    wait_for_selector=".article-card",  # Wait for specific elements
+    timeout=30000                      # Max 30 seconds total
 )
 
 # Use it in your crawl
@@ -144,17 +160,29 @@ async with AsyncWebCrawler() as crawler:
 ### The Three-Layer Scoring System
 
 ```python
-from crawl4ai import LinkPreviewConfig, CrawlerRunConfig, CacheMode
+from crawl4ai import LinkPreviewConfig
 
 # Configure intelligent link analysis
 link_config = LinkPreviewConfig(
+    # What to analyze
     include_internal=True,
-    include_external=False,
+    include_external=True,
-    max_links=10,
+    max_links=100,              # Analyze top 100 links
-    concurrency=5,
+    
-    query="python tutorial",  # For contextual scoring
+    # Relevance scoring
-    score_threshold=0.3,
+    query="machine learning tutorials",  # Your interest
-    verbose=True
+    score_threshold=0.3,        # Minimum relevance score
+    
+    # Performance
+    concurrent_requests=10,     # Parallel processing
+    timeout_per_link=5000,      # 5s per link
+    
+    # Advanced scoring weights
+    scoring_weights={
+        "intrinsic": 0.3,       # Link quality indicators
+        "contextual": 0.5,      # Relevance to query
+        "popularity": 0.2       # Link prominence
+    }
 )
 
 # Use in your crawl
@@ -162,51 +190,35 @@ result = await crawler.arun(
     "https://tech-blog.example.com",
     config=CrawlerRunConfig(
         link_preview_config=link_config,
-        score_links=True,  # Enable intrinsic scoring
+        score_links=True
-        cache_mode=CacheMode.BYPASS
     )
 )
 
 # Access scored and sorted links
-if result.success and result.links:
+for link in result.links["internal"][:10]:  # Top 10 internal links
-# Get scored links
+    print(f"Score: {link['total_score']:.3f}")
-internal_links = result.links.get("internal", [])
+    print(f"  Intrinsic: {link['intrinsic_score']:.1f}/10")  # Position, attributes
-scored_links = [l for l in internal_links if l.get("total_score")]
+    print(f"  Contextual: {link['contextual_score']:.1f}/1")  # Relevance to query
-scored_links.sort(key=lambda x: x.get("total_score", 0), reverse=True)
+    print(f"  URL: {link['href']}")
-
+    print(f"  Title: {link['head_data']['title']}")
-# Create a scoring table
+    print(f"  Description: {link['head_data']['meta']['description'][:100]}...")
-table = Table(title="Link Scoring Results", box=box.ROUNDED)
-table.add_column("Link Text", style="cyan", width=40)
-table.add_column("Intrinsic Score", justify="center")
-table.add_column("Contextual Score", justify="center")
-table.add_column("Total Score", justify="center", style="bold green")
-
-for link in scored_links[:5]:
-    text = link.get('text', 'No text')[:40]
-    table.add_row(
-        text,
-        f"{link.get('intrinsic_score', 0):.1f}/10",
-        f"{link.get('contextual_score', 0):.2f}/1",
-        f"{link.get('total_score', 0):.3f}"
-    )
-
-console.print(table)
 ```
 
 **Scoring Components:**
 
-1. **Intrinsic Score**: Based on link quality indicators
+1. **Intrinsic Score (0-10)**: 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
+2. **Contextual Score (0-1)**: Relevance to your query
+   - Semantic similarity using embeddings
    - Keyword matching in link text and title
    - Meta description analysis
    - Content preview scoring
 
-3. **Total Score**: Combined score for final ranking
+3. **Total Score**: Weighted combination for final ranking
 
 **Expected Real-World Impact:**
 - **Research Efficiency**: Find relevant papers 10x faster by following only high-score links
@@ -228,53 +240,53 @@ from crawl4ai import AsyncUrlSeeder, SeedingConfig
 # Basic discovery - find all product pages
 seeder_config = SeedingConfig(
     # Discovery sources
-    source="cc+sitemap",        # Sitemap + Common Crawl
+    source="sitemap+cc",        # Sitemap + Common Crawl
     
     # Filtering
     pattern="*/product/*",      # URL pattern matching
+    ignore_patterns=["*/reviews/*", "*/questions/*"],
     
     # Validation
     live_check=True,           # Verify URLs are alive
-    max_urls=50,             # Stop at 50 URLs
+    max_urls=5000,             # Stop at 5000 URLs
     
     # Performance  
-    concurrency=100,           # Maximum concurrent requests for live checks/head extraction
+    concurrency=100,           # Parallel requests
-    hits_per_sec=10           # Rate limit in requests per second to avoid overwhelming servers
+    hits_per_sec=10           # Rate limiting
 )
 
-async with AsyncUrlSeeder() as seeder:
+seeder = AsyncUrlSeeder(seeder_config)
-    console.print("Discovering URLs from Python docs...")
+urls = await seeder.discover("https://shop.example.com")
-    urls = await seeder.urls("docs.python.org", seeding_config)
-    console.print(f"\n✓ Discovered {len(urls)} URLs")
 
 # Advanced: Relevance-based discovery
 research_config = SeedingConfig(
-    source="sitemap+cc",       # Sitemap + Common Crawl
+    source="crawl+sitemap",    # Deep crawl + sitemap
     pattern="*/blog/*",        # Blog posts only
     
     # Content relevance
     extract_head=True,         # Get meta tags
     query="quantum computing tutorials",
-    scoring_method="bm25",     # BM25 scoring method
+    scoring_method="bm25",     # Or "semantic" (coming soon)
     score_threshold=0.4,       # High relevance only
     
     # Smart filtering
     filter_nonsense_urls=True,  # Remove .xml, .txt, etc.
+    min_content_length=500,     # Skip thin content
     
     force=True                 # Bypass cache
 )
 
 # Discover with progress tracking
 discovered = []
-async with AsyncUrlSeeder() as seeder:
+async for batch in seeder.discover_iter("https://physics-blog.com", research_config):
-    discovered = await seeder.urls("https://physics-blog.com", research_config)
+    discovered.extend(batch)
-    console.print(f"\n✓ Discovered {len(discovered)} URLs")
+    print(f"Found {len(discovered)} relevant URLs so far...")
 
 # Results include scores and metadata
 for url_data in discovered[:5]:
     print(f"URL: {url_data['url']}")
-    print(f"Score: {url_data['relevance_score']:.3f}")
+    print(f"Score: {url_data['score']:.3f}")
-    print(f"Title: {url_data['head_data']['title']}")
+    print(f"Title: {url_data['title']}")
 ```
 
 **Discovery Methods:**
@@ -297,18 +309,35 @@ This release includes significant performance improvements through optimized res
 ### What We Optimized
 
 ```python
-# Optimized crawling with v0.7.0 improvements
+# Before v0.7.0 (slow)
 results = []
 for url in urls:
-    result = await crawler.arun(
+    result = await crawler.arun(url)
-        url,
-        config=CrawlerRunConfig(
-            # Performance optimizations
-            wait_until="domcontentloaded",  # Faster than networkidle
-            cache_mode=CacheMode.ENABLED    # Enable caching
-        )
-    )
     results.append(result)
+
+# After v0.7.0 (fast)
+# Automatic batching and connection pooling
+results = await crawler.arun_batch(
+    urls,
+    config=CrawlerRunConfig(
+        # New performance options
+        batch_size=10,              # Process 10 URLs concurrently
+        reuse_browser=True,         # Keep browser warm
+        eager_loading=False,        # Load only what's needed
+        streaming_extraction=True,  # Stream large extractions
+        
+        # Optimized defaults
+        wait_until="domcontentloaded",  # Faster than networkidle
+        exclude_external_resources=True, # Skip third-party assets
+        block_ads=True                  # Ad blocking built-in
+    )
+)
+
+# Memory-efficient streaming for large crawls
+async for result in crawler.arun_stream(large_url_list):
+    # Process results as they complete
+    await process_result(result)
+    # Memory is freed after each iteration
 ```
 
 **Performance Gains:**
@@ -318,6 +347,24 @@ for url in urls:
 - **Memory Usage**: 60% reduction with streaming processing
 - **Concurrent Crawls**: Handle 5x more parallel requests
 
+## 📄 PDF Support
+
+PDF extraction is now natively supported in Crawl4AI.
+
+```python
+# Extract data from PDF documents
+result = await crawler.arun(
+    "https://example.com/report.pdf",
+    config=CrawlerRunConfig(
+        pdf_extraction=True,
+        extraction_strategy=JsonCssExtractionStrategy({
+            # Works on converted PDF structure
+            "title": {"selector": "h1", "type": "text"},
+            "sections": {"selector": "h2", "type": "list"}
+        })
+    )
+)
+```
 
 ## 🔧 Important Changes
 

docs/md_v2/advanced/session-management.md

@@ -49,75 +49,46 @@ from crawl4ai import JsonCssExtractionStrategy
 from crawl4ai.cache_context import CacheMode
 
 async def crawl_dynamic_content():
-    url = "https://github.com/microsoft/TypeScript/commits/main"
+    async with AsyncWebCrawler() as crawler:
-    session_id = "wait_for_session"
+        session_id = "github_commits_session"
-    all_commits = []
+        url = "https://github.com/microsoft/TypeScript/commits/main"
+        all_commits = []
 
-    js_next_page = """
+        # Define extraction schema
-    const commits = document.querySelectorAll('li[data-testid="commit-row-item"] h4');
+        schema = {
-    if (commits.length > 0) {
+            "name": "Commit Extractor",
-        window.lastCommit = commits[0].textContent.trim();
+            "baseSelector": "li.Box-sc-g0xbh4-0",
-    }
+            "fields": [{
-    const button = document.querySelector('a[data-testid="pagination-next-button"]');
+                "name": "title", "selector": "h4.markdown-title", "type": "text"
-    if (button) {button.click(); console.log('button clicked') }
+            }],
-    """
+        }
+        extraction_strategy = JsonCssExtractionStrategy(schema)
 
-    wait_for = """() => {
+        # JavaScript and wait configurations
-        const commits = document.querySelectorAll('li[data-testid="commit-row-item"] h4');
+        js_next_page = """document.querySelector('a[data-testid="pagination-next-button"]').click();"""
-        if (commits.length === 0) return false;
+        wait_for = """() => document.querySelectorAll('li.Box-sc-g0xbh4-0').length > 0"""
-        const firstCommit = commits[0].textContent.trim();
+
-        return firstCommit !== window.lastCommit;
+        # Crawl multiple pages
-    }"""
-    
-    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):
-            crawler_config = CrawlerRunConfig(
+            config = CrawlerRunConfig(
+                url=url,
                 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(url=url, config=crawler_config)
+            result = await crawler.arun(config=config)
-            
+            if result.success:
-            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/virtual-scroll.md

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

docs/md_v2/blog/releases/0.7.0.md

@@ -10,8 +10,9 @@ Today I'm releasing Crawl4AI v0.7.0—the Adaptive Intelligence Update. This rel
 
 - **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
+- **Link Preview with 3-Layer Scoring**: Intelligent link analysis and prioritization
 - **Async URL Seeder**: Discover thousands of URLs in seconds with intelligent filtering
+- **PDF Parsing**: Extract data from PDF documents
 - **Performance Optimizations**: Significant speed and memory improvements
 
 ## 🧠 Adaptive Crawling: Intelligence Through Pattern Learning
@@ -29,34 +30,44 @@ The Adaptive Crawler maintains a persistent state for each domain, tracking:
 - Extraction confidence scores
 
 ```python
-from crawl4ai import AsyncWebCrawler, AdaptiveCrawler, AdaptiveConfig
+from crawl4ai import AdaptiveCrawler, AdaptiveConfig, CrawlState
 
-# Initialize with custom adaptive parameters
+# Initialize with custom learning parameters
 config = AdaptiveConfig(
-    confidence_threshold=0.7,    # Min confidence to stop crawling
+    confidence_threshold=0.7,    # Min confidence to use learned patterns
-    max_depth=5,                # Maximum crawl depth
+    max_history=100,            # Remember last 100 crawls per domain
-    max_pages=20,               # Maximum number of pages to crawl
+    learning_rate=0.2,          # How quickly to adapt to changes
-    top_k_links=3,              # Number of top links to follow per page
+    patterns_per_page=3,        # Patterns to learn per page type
-    strategy="statistical",     # 'statistical' or 'embedding'
+    extraction_strategy='css'   # 'css' or 'xpath'
-    coverage_weight=0.4,        # Weight for coverage in confidence calculation
-    consistency_weight=0.3,     # Weight for consistency in confidence calculation
-    saturation_weight=0.3       # Weight for saturation in confidence calculation
 )
 
-# Initialize adaptive crawler with web crawler
+adaptive_crawler = AdaptiveCrawler(config)
+
+# First crawl - crawler learns the structure
 async with AsyncWebCrawler() as crawler:
-    adaptive_crawler = AdaptiveCrawler(crawler, config)
+    result = await crawler.arun(
-    
+        "https://news.example.com/article/12345",
-    # Crawl and learn patterns
+        config=CrawlerRunConfig(
-    state = await adaptive_crawler.digest(
+            adaptive_config=config,
-        start_url="https://news.example.com/article/12345",
+            extraction_hints={  # Optional hints to speed up learning
-        query="latest news articles and content"
+                "title": "article h1",
+                "content": "article .body-content"
+            }
+        )
     )
     
-    # Access results and confidence
+    # Crawler identifies and stores patterns
-    print(f"Confidence Level: {adaptive_crawler.confidence:.0%}")
+    if result.success:
-    print(f"Pages Crawled: {len(state.crawled_urls)}")
+        state = adaptive_crawler.get_state("news.example.com")
-    print(f"Knowledge Base: {len(adaptive_crawler.state.knowledge_base)} documents")
+        print(f"Learned {len(state.patterns)} patterns")
+        print(f"Confidence: {state.avg_confidence:.2%}")
+
+# Subsequent crawls - uses learned patterns
+result2 = await crawler.arun(
+    "https://news.example.com/article/67890",
+    config=CrawlerRunConfig(adaptive_config=config)
+)
+# Automatically extracts using learned patterns!
 ```
 
 **Expected Real-World Impact:**
@@ -81,7 +92,9 @@ 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
+    wait_after_scroll=1.0,             # Let content load
+    capture_method="incremental",       # Capture new content on each scroll
+    deduplicate=True                   # Remove duplicate elements
 )
 
 # For e-commerce product grids (Instagram style)
@@ -89,7 +102,8 @@ 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
+    wait_after_scroll=1.5,             # Images need time
+    stop_on_no_change=True            # Smart stopping
 )
 
 # For news feeds with lazy loading
@@ -97,7 +111,9 @@ 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
+    wait_after_scroll=0.5,
+    wait_for_selector=".article-card",  # Wait for specific elements
+    timeout=30000                      # Max 30 seconds total
 )
 
 # Use it in your crawl
@@ -144,17 +160,29 @@ async with AsyncWebCrawler() as crawler:
 ### The Three-Layer Scoring System
 
 ```python
-from crawl4ai import LinkPreviewConfig, CrawlerRunConfig, CacheMode
+from crawl4ai import LinkPreviewConfig
 
 # Configure intelligent link analysis
 link_config = LinkPreviewConfig(
+    # What to analyze
     include_internal=True,
-    include_external=False,
+    include_external=True,
-    max_links=10,
+    max_links=100,              # Analyze top 100 links
-    concurrency=5,
+    
-    query="python tutorial",  # For contextual scoring
+    # Relevance scoring
-    score_threshold=0.3,
+    query="machine learning tutorials",  # Your interest
-    verbose=True
+    score_threshold=0.3,        # Minimum relevance score
+    
+    # Performance
+    concurrent_requests=10,     # Parallel processing
+    timeout_per_link=5000,      # 5s per link
+    
+    # Advanced scoring weights
+    scoring_weights={
+        "intrinsic": 0.3,       # Link quality indicators
+        "contextual": 0.5,      # Relevance to query
+        "popularity": 0.2       # Link prominence
+    }
 )
 
 # Use in your crawl
@@ -162,51 +190,35 @@ result = await crawler.arun(
     "https://tech-blog.example.com",
     config=CrawlerRunConfig(
         link_preview_config=link_config,
-        score_links=True,  # Enable intrinsic scoring
+        score_links=True
-        cache_mode=CacheMode.BYPASS
     )
 )
 
 # Access scored and sorted links
-if result.success and result.links:
+for link in result.links["internal"][:10]:  # Top 10 internal links
-# Get scored links
+    print(f"Score: {link['total_score']:.3f}")
-internal_links = result.links.get("internal", [])
+    print(f"  Intrinsic: {link['intrinsic_score']:.1f}/10")  # Position, attributes
-scored_links = [l for l in internal_links if l.get("total_score")]
+    print(f"  Contextual: {link['contextual_score']:.1f}/1")  # Relevance to query
-scored_links.sort(key=lambda x: x.get("total_score", 0), reverse=True)
+    print(f"  URL: {link['href']}")
-
+    print(f"  Title: {link['head_data']['title']}")
-# Create a scoring table
+    print(f"  Description: {link['head_data']['meta']['description'][:100]}...")
-table = Table(title="Link Scoring Results", box=box.ROUNDED)
-table.add_column("Link Text", style="cyan", width=40)
-table.add_column("Intrinsic Score", justify="center")
-table.add_column("Contextual Score", justify="center")
-table.add_column("Total Score", justify="center", style="bold green")
-
-for link in scored_links[:5]:
-    text = link.get('text', 'No text')[:40]
-    table.add_row(
-        text,
-        f"{link.get('intrinsic_score', 0):.1f}/10",
-        f"{link.get('contextual_score', 0):.2f}/1",
-        f"{link.get('total_score', 0):.3f}"
-    )
-
-console.print(table)
 ```
 
 **Scoring Components:**
 
-1. **Intrinsic Score**: Based on link quality indicators
+1. **Intrinsic Score (0-10)**: 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
+2. **Contextual Score (0-1)**: Relevance to your query
+   - Semantic similarity using embeddings
    - Keyword matching in link text and title
    - Meta description analysis
    - Content preview scoring
 
-3. **Total Score**: Combined score for final ranking
+3. **Total Score**: Weighted combination for final ranking
 
 **Expected Real-World Impact:**
 - **Research Efficiency**: Find relevant papers 10x faster by following only high-score links
@@ -228,53 +240,53 @@ from crawl4ai import AsyncUrlSeeder, SeedingConfig
 # Basic discovery - find all product pages
 seeder_config = SeedingConfig(
     # Discovery sources
-    source="cc+sitemap",        # Sitemap + Common Crawl
+    source="sitemap+cc",        # Sitemap + Common Crawl
     
     # Filtering
     pattern="*/product/*",      # URL pattern matching
+    ignore_patterns=["*/reviews/*", "*/questions/*"],
     
     # Validation
     live_check=True,           # Verify URLs are alive
-    max_urls=50,             # Stop at 50 URLs
+    max_urls=5000,             # Stop at 5000 URLs
     
     # Performance  
-    concurrency=100,           # Maximum concurrent requests for live checks/head extraction
+    concurrency=100,           # Parallel requests
-    hits_per_sec=10           # Rate limit in requests per second to avoid overwhelming servers
+    hits_per_sec=10           # Rate limiting
 )
 
-async with AsyncUrlSeeder() as seeder:
+seeder = AsyncUrlSeeder(seeder_config)
-    console.print("Discovering URLs from Python docs...")
+urls = await seeder.discover("https://shop.example.com")
-    urls = await seeder.urls("docs.python.org", seeding_config)
-    console.print(f"\n✓ Discovered {len(urls)} URLs")
 
 # Advanced: Relevance-based discovery
 research_config = SeedingConfig(
-    source="sitemap+cc",       # Sitemap + Common Crawl
+    source="crawl+sitemap",    # Deep crawl + sitemap
     pattern="*/blog/*",        # Blog posts only
     
     # Content relevance
     extract_head=True,         # Get meta tags
     query="quantum computing tutorials",
-    scoring_method="bm25",     # BM25 scoring method
+    scoring_method="bm25",     # Or "semantic" (coming soon)
     score_threshold=0.4,       # High relevance only
     
     # Smart filtering
     filter_nonsense_urls=True,  # Remove .xml, .txt, etc.
+    min_content_length=500,     # Skip thin content
     
     force=True                 # Bypass cache
 )
 
 # Discover with progress tracking
 discovered = []
-async with AsyncUrlSeeder() as seeder:
+async for batch in seeder.discover_iter("https://physics-blog.com", research_config):
-    discovered = await seeder.urls("https://physics-blog.com", research_config)
+    discovered.extend(batch)
-    console.print(f"\n✓ Discovered {len(discovered)} URLs")
+    print(f"Found {len(discovered)} relevant URLs so far...")
 
 # Results include scores and metadata
 for url_data in discovered[:5]:
     print(f"URL: {url_data['url']}")
-    print(f"Score: {url_data['relevance_score']:.3f}")
+    print(f"Score: {url_data['score']:.3f}")
-    print(f"Title: {url_data['head_data']['title']}")
+    print(f"Title: {url_data['title']}")
 ```
 
 **Discovery Methods:**
@@ -297,18 +309,35 @@ This release includes significant performance improvements through optimized res
 ### What We Optimized
 
 ```python
-# Optimized crawling with v0.7.0 improvements
+# Before v0.7.0 (slow)
 results = []
 for url in urls:
-    result = await crawler.arun(
+    result = await crawler.arun(url)
-        url,
-        config=CrawlerRunConfig(
-            # Performance optimizations
-            wait_until="domcontentloaded",  # Faster than networkidle
-            cache_mode=CacheMode.ENABLED    # Enable caching
-        )
-    )
     results.append(result)
+
+# After v0.7.0 (fast)
+# Automatic batching and connection pooling
+results = await crawler.arun_batch(
+    urls,
+    config=CrawlerRunConfig(
+        # New performance options
+        batch_size=10,              # Process 10 URLs concurrently
+        reuse_browser=True,         # Keep browser warm
+        eager_loading=False,        # Load only what's needed
+        streaming_extraction=True,  # Stream large extractions
+        
+        # Optimized defaults
+        wait_until="domcontentloaded",  # Faster than networkidle
+        exclude_external_resources=True, # Skip third-party assets
+        block_ads=True                  # Ad blocking built-in
+    )
+)
+
+# Memory-efficient streaming for large crawls
+async for result in crawler.arun_stream(large_url_list):
+    # Process results as they complete
+    await process_result(result)
+    # Memory is freed after each iteration
 ```
 
 **Performance Gains:**
@@ -318,6 +347,24 @@ for url in urls:
 - **Memory Usage**: 60% reduction with streaming processing
 - **Concurrent Crawls**: Handle 5x more parallel requests
 
+## 📄 PDF Support
+
+PDF extraction is now natively supported in Crawl4AI.
+
+```python
+# Extract data from PDF documents
+result = await crawler.arun(
+    "https://example.com/report.pdf",
+    config=CrawlerRunConfig(
+        pdf_extraction=True,
+        extraction_strategy=JsonCssExtractionStrategy({
+            # Works on converted PDF structure
+            "title": {"selector": "h1", "type": "text"},
+            "sections": {"selector": "h2", "type": "list"}
+        })
+    )
+)
+```
 
 ## 🔧 Important Changes
 

docs/md_v2/core/adaptive-crawling.md

@@ -35,7 +35,7 @@ from crawl4ai import AsyncWebCrawler, AdaptiveCrawler
 
 async def main():
     async with AsyncWebCrawler() as crawler:
-        # Create an adaptive crawler (config is optional)
+        # Create an adaptive crawler
         adaptive = AdaptiveCrawler(crawler)
         
         # Start crawling with a query
@@ -59,13 +59,13 @@ async def main():
 from crawl4ai import AdaptiveConfig
 
 config = AdaptiveConfig(
-    confidence_threshold=0.8,    # Stop when 80% confident (default: 0.7)
+    confidence_threshold=0.7,    # Stop when 70% confident (default: 0.8)
-    max_pages=30,               # Maximum pages to crawl (default: 20)
+    max_pages=20,               # Maximum pages to crawl (default: 50)
-    top_k_links=5,              # Links to follow per page (default: 3)
+    top_k_links=3,              # Links to follow per page (default: 5)
     min_gain_threshold=0.05     # Minimum expected gain to continue (default: 0.1)
 )
 
-adaptive = AdaptiveCrawler(crawler, config)
+adaptive = AdaptiveCrawler(crawler, config=config)
 ```
 
 ## Crawling Strategies
@@ -198,8 +198,8 @@ if result.metrics.get('is_irrelevant', False):
 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
+- **0.6-0.8**: Good coverage, can answer most queries
-- **0.7-1.0**: Excellent coverage, comprehensive information
+- **0.8-1.0**: Excellent coverage, comprehensive information
 
 ### Statistics Display
 
@@ -257,9 +257,9 @@ new_adaptive.import_knowledge_base("knowledge_base.jsonl")
 - Avoid overly broad queries
 
 ### 2. Threshold Tuning
-- Start with default (0.7) for general use
+- Start with default (0.8) for general use
-- Lower to 0.5-0.6 for exploratory crawling
+- Lower to 0.6-0.7 for exploratory crawling
-- Raise to 0.8+ for exhaustive coverage
+- Raise to 0.9+ for exhaustive coverage
 
 ### 3. Performance Optimization
 - Use appropriate `max_pages` limits

docs/md_v2/core/url-seeding.md

@@ -137,7 +137,7 @@ async def smart_blog_crawler():
             word_count_threshold=300  # Only substantial articles
         )
         
-        # Extract URLs and crawl them
+        # Extract URLs and stream results as they come
         tutorial_urls = [t["url"] for t in tutorials[:10]]
         results = await crawler.arun_many(tutorial_urls, config=config)
         
@@ -231,7 +231,7 @@ Common Crawl is a massive public dataset that regularly crawls the entire web. I
 
 ```python
 # Use both sources
-config = SeedingConfig(source="sitemap+cc")
+config = SeedingConfig(source="cc+sitemap")
 urls = await seeder.urls("example.com", config)
 ```
 
@@ -241,13 +241,13 @@ The `SeedingConfig` object is your control panel. Here's everything you can conf
 
 | Parameter | Type | Default | Description |
 |-----------|------|---------|-------------|
-| `source` | str | "sitemap+cc" | URL source: "cc" (Common Crawl), "sitemap", or "sitemap+cc" |
+| `source` | str | "cc" | URL source: "cc" (Common Crawl), "sitemap", or "cc+sitemap" |
 | `pattern` | str | "*" | URL pattern filter (e.g., "*/blog/*", "*.html") |
 | `extract_head` | bool | False | Extract metadata from page `<head>` |
 | `live_check` | bool | False | Verify URLs are accessible |
 | `max_urls` | int | -1 | Maximum URLs to return (-1 = unlimited) |
 | `concurrency` | int | 10 | Parallel workers for fetching |
-| `hits_per_sec` | int | 5 | Rate limit for requests |
+| `hits_per_sec` | int | None | Rate limit for requests |
 | `force` | bool | False | Bypass cache, fetch fresh data |
 | `verbose` | bool | False | Show detailed progress |
 | `query` | str | None | Search query for BM25 scoring |
@@ -522,7 +522,7 @@ urls = await seeder.urls("docs.example.com", config)
 ```python
 # Find specific products
 config = SeedingConfig(
-    source="sitemap+cc",  # Use both sources
+    source="cc+sitemap",  # Use both sources
     extract_head=True,
     query="wireless headphones noise canceling",
     scoring_method="bm25",
@@ -782,7 +782,7 @@ class ResearchAssistant:
         
         # Step 1: Discover relevant URLs
         config = SeedingConfig(
-            source="sitemap+cc",     # Maximum coverage
+            source="cc+sitemap",     # Maximum coverage
             extract_head=True,       # Get metadata
             query=topic,             # Research topic
             scoring_method="bm25",   # Smart scoring
@@ -832,8 +832,7 @@ class ResearchAssistant:
             # Extract URLs and crawl all articles
             article_urls = [article['url'] for article in top_articles]
             results = []
-            crawl_results = await crawler.arun_many(article_urls, config=config)
+            async for result in await crawler.arun_many(article_urls, config=config):
-            async for result in crawl_results:
                 if result.success:
                     results.append({
                         'url': result.url,
@@ -934,10 +933,10 @@ config = SeedingConfig(concurrency=10, hits_per_sec=5)
 # When crawling many URLs
 async with AsyncWebCrawler() as crawler:
     # Assuming urls is a list of URL strings
-    crawl_results = await crawler.arun_many(urls, config=config)
+    results = await crawler.arun_many(urls, config=config)
     
     # Process as they arrive
-    async for result in crawl_results:
+    async for result in results:
         process_immediately(result)  # Don't wait for all
 ```
 
@@ -1021,7 +1020,7 @@ config = SeedingConfig(
 
 # E-commerce product discovery
 config = SeedingConfig(
-    source="sitemap+cc",
+    source="cc+sitemap",
     pattern="*/product/*",
     extract_head=True,
     live_check=True

test_stealth_compatibility.py

@@ -0,0 +1,141 @@
+#!/usr/bin/env python3
+"""
+Test suite for playwright-stealth backward compatibility.
+Tests that stealth functionality works automatically without user configuration.
+"""
+
+import pytest
+import asyncio
+from unittest.mock import Mock, patch, MagicMock
+
+
+class TestPlaywrightStealthCompatibility:
+    """Test playwright-stealth backward compatibility with transparent operation"""
+
+    def test_api_detection_works(self):
+        """Test that API detection works correctly"""
+        from crawl4ai.async_crawler_strategy import STEALTH_NEW_API
+        # The value depends on which version is installed, but should not be undefined
+        assert STEALTH_NEW_API is not None or STEALTH_NEW_API is False or STEALTH_NEW_API is None
+
+    @pytest.mark.asyncio
+    @patch('crawl4ai.async_crawler_strategy.STEALTH_NEW_API', True)
+    @patch('crawl4ai.async_crawler_strategy.Stealth')
+    async def test_apply_stealth_new_api(self, mock_stealth_class):
+        """Test stealth application with new API works transparently"""
+        from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+        
+        # Setup mock
+        mock_stealth_instance = Mock()
+        mock_stealth_instance.apply_stealth_async = Mock()
+        mock_stealth_class.return_value = mock_stealth_instance
+        
+        # Create strategy instance
+        strategy = AsyncPlaywrightCrawlerStrategy()
+        
+        # Mock page
+        mock_page = Mock()
+        
+        # Test the method - should work transparently
+        await strategy._apply_stealth(mock_page)
+        
+        # Verify new API was used
+        mock_stealth_class.assert_called_once()
+        mock_stealth_instance.apply_stealth_async.assert_called_once_with(mock_page)
+
+    @pytest.mark.asyncio
+    @patch('crawl4ai.async_crawler_strategy.STEALTH_NEW_API', False)
+    async def test_apply_stealth_legacy_api(self):
+        """Test stealth application with legacy API works transparently"""
+        from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+        
+        # Mock stealth_async function by setting it as a module attribute
+        mock_stealth_async = Mock()
+        mock_stealth_async.return_value = None
+        
+        # Import the module to add the mock function
+        import crawl4ai.async_crawler_strategy
+        crawl4ai.async_crawler_strategy.stealth_async = mock_stealth_async
+        
+        try:
+            # Create strategy instance
+            strategy = AsyncPlaywrightCrawlerStrategy()
+            
+            # Mock page
+            mock_page = Mock()
+            
+            # Test the method - should work transparently
+            await strategy._apply_stealth(mock_page)
+            
+            # Verify legacy API was used
+            mock_stealth_async.assert_called_once_with(mock_page)
+        finally:
+            # Clean up
+            if hasattr(crawl4ai.async_crawler_strategy, 'stealth_async'):
+                delattr(crawl4ai.async_crawler_strategy, 'stealth_async')
+
+    @pytest.mark.asyncio
+    @patch('crawl4ai.async_crawler_strategy.STEALTH_NEW_API', None)
+    async def test_apply_stealth_no_library(self):
+        """Test stealth application when no stealth library is available"""
+        from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+        
+        # Create strategy instance
+        strategy = AsyncPlaywrightCrawlerStrategy()
+        
+        # Mock page
+        mock_page = Mock()
+        
+        # Test the method - should work transparently even without stealth
+        await strategy._apply_stealth(mock_page)
+        
+        # Should complete without error even when no stealth is available
+
+    @pytest.mark.asyncio
+    @patch('crawl4ai.async_crawler_strategy.STEALTH_NEW_API', True)
+    @patch('crawl4ai.async_crawler_strategy.Stealth')
+    async def test_stealth_error_handling(self, mock_stealth_class):
+        """Test that stealth errors are handled gracefully without breaking crawling"""
+        from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+        
+        # Setup mock to raise an error
+        mock_stealth_instance = Mock()
+        mock_stealth_instance.apply_stealth_async = Mock(side_effect=Exception("Stealth failed"))
+        mock_stealth_class.return_value = mock_stealth_instance
+        
+        # Create strategy instance
+        strategy = AsyncPlaywrightCrawlerStrategy()
+        
+        # Mock page
+        mock_page = Mock()
+        
+        # Test the method - should not raise an error, continue silently
+        await strategy._apply_stealth(mock_page)
+        
+        # Should complete without raising the stealth error
+
+    def test_strategy_creation_without_config(self):
+        """Test that strategy can be created without any stealth configuration"""
+        from crawl4ai.async_crawler_strategy import AsyncPlaywrightCrawlerStrategy
+        
+        # Should work without any stealth-related parameters
+        strategy = AsyncPlaywrightCrawlerStrategy()
+        assert strategy is not None
+        assert hasattr(strategy, '_apply_stealth')
+
+    def test_browser_config_works_without_stealth_param(self):
+        """Test that BrowserConfig works without stealth parameter"""
+        from crawl4ai.async_configs import BrowserConfig
+        
+        # Should work without stealth parameter
+        config = BrowserConfig()
+        assert config is not None
+        
+        # Should also work with other parameters
+        config = BrowserConfig(headless=False, browser_type="firefox")
+        assert config.headless == False
+        assert config.browser_type == "firefox"
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])