crawl4ai/docs/md_v2/blog/releases/0.5.0.md

# Crawl4AI v0.5.0 Release Notes

**Release Theme: Power, Flexibility, and Scalability**

Crawl4AI v0.5.0 is a major release focused on significantly enhancing the
library's power, flexibility, and scalability. Key improvements include a new
**deep crawling** system, a **memory-adaptive dispatcher** for handling
large-scale crawls, **multiple crawling strategies** (including a fast HTTP-only
crawler), **Docker** deployment options, and a powerful **command-line interface
(CLI)**. This release also includes numerous bug fixes, performance
optimizations, and documentation updates.

**Important Note:** This release contains several **breaking changes**. Please
review the "Breaking Changes" section carefully and update your code
accordingly.

## Key Features

### 1. Deep Crawling

Crawl4AI now supports deep crawling, allowing you to explore websites beyond the
initial URLs. This is controlled by the `deep_crawl_strategy` parameter in
`CrawlerRunConfig`. Several strategies are available:

- **`BFSDeepCrawlStrategy` (Breadth-First Search):** Explores the website level
  by level. (Default)
- **`DFSDeepCrawlStrategy` (Depth-First Search):** Explores each branch as
  deeply as possible before backtracking.
- **`BestFirstCrawlingStrategy`:** Uses a scoring function to prioritize which
  URLs to crawl next.

```python
import time
from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, BFSDeepCrawlStrategy
from crawl4ai.content_scraping_strategy import LXMLWebScrapingStrategy
from crawl4ai.deep_crawling import DomainFilter, ContentTypeFilter, FilterChain, URLPatternFilter, KeywordRelevanceScorer, BestFirstCrawlingStrategy
import asyncio

# Create a filter chain to filter urls based on patterns, domains and content type
filter_chain = FilterChain(
    [
        DomainFilter(
            allowed_domains=["docs.crawl4ai.com"],
            blocked_domains=["old.docs.crawl4ai.com"],
        ),
        URLPatternFilter(patterns=["*core*", "*advanced*"],),
        ContentTypeFilter(allowed_types=["text/html"]),
    ]
)

# Create a keyword scorer that prioritises the pages with certain keywords first
keyword_scorer = KeywordRelevanceScorer(
    keywords=["crawl", "example", "async", "configuration"], weight=0.7
)

# Set up the configuration
deep_crawl_config = CrawlerRunConfig(
    deep_crawl_strategy=BestFirstCrawlingStrategy(
        max_depth=2,
        include_external=False,
        filter_chain=filter_chain,
        url_scorer=keyword_scorer,
    ),
    scraping_strategy=LXMLWebScrapingStrategy(),
    stream=True,
    verbose=True,
)

async def main():
    async with AsyncWebCrawler() as crawler:
        start_time = time.perf_counter()
        results = []
        async for result in await crawler.arun(url="https://docs.crawl4ai.com", config=deep_crawl_config):
            print(f"Crawled: {result.url} (Depth: {result.metadata['depth']}), score: {result.metadata['score']:.2f}")
            results.append(result)
        duration = time.perf_counter() - start_time
        print(f"\n✅ Crawled {len(results)} high-value pages in {duration:.2f} seconds")

asyncio.run(main())
```

**Breaking Change:** The `max_depth` parameter is now part of `CrawlerRunConfig`
and controls the _depth_ of the crawl, not the number of concurrent crawls. The
`arun()` and `arun_many()` methods are now decorated to handle deep crawling
strategies. Imports for deep crawling strategies have changed. See the
[Deep Crawling documentation](../../core/deep-crawling.md) for more details.

### 2. Memory-Adaptive Dispatcher

The new `MemoryAdaptiveDispatcher` dynamically adjusts concurrency based on
available system memory and includes built-in rate limiting. This prevents
out-of-memory errors and avoids overwhelming target websites.

```python
from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, MemoryAdaptiveDispatcher
import asyncio

# Configure the dispatcher (optional, defaults are used if not provided)
dispatcher = MemoryAdaptiveDispatcher(
    memory_threshold_percent=80.0,  # Pause if memory usage exceeds 80%
    check_interval=0.5,  # Check memory every 0.5 seconds
)

async def batch_mode():
    async with AsyncWebCrawler() as crawler:
        results = await crawler.arun_many(
            urls=["https://docs.crawl4ai.com", "https://github.com/unclecode/crawl4ai"],
            config=CrawlerRunConfig(stream=False),  # Batch mode
            dispatcher=dispatcher,
        )
        for result in results:
            print(f"Crawled: {result.url} with status code: {result.status_code}")

async def stream_mode():
    async with AsyncWebCrawler() as crawler:
        # OR, for streaming:
        async for result in await crawler.arun_many(
            urls=["https://docs.crawl4ai.com", "https://github.com/unclecode/crawl4ai"],
            config=CrawlerRunConfig(stream=True),
            dispatcher=dispatcher,
        ):
            print(f"Crawled: {result.url} with status code: {result.status_code}")

print("Dispatcher in batch mode:")
asyncio.run(batch_mode())
print("-" * 50)
print("Dispatcher in stream mode:")
asyncio.run(stream_mode())
```

**Breaking Change:** `AsyncWebCrawler.arun_many()` now uses
`MemoryAdaptiveDispatcher` by default. Existing code that relied on unbounded
concurrency may require adjustments.

### 3. Multiple Crawling Strategies (Playwright and HTTP)

Crawl4AI now offers two crawling strategies:

- **`AsyncPlaywrightCrawlerStrategy` (Default):** Uses Playwright for
  browser-based crawling, supporting JavaScript rendering and complex
  interactions.
- **`AsyncHTTPCrawlerStrategy`:** A lightweight, fast, and memory-efficient
  HTTP-only crawler. Ideal for simple scraping tasks where browser rendering is
  unnecessary.

```python
from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, HTTPCrawlerConfig
from crawl4ai.async_crawler_strategy import AsyncHTTPCrawlerStrategy
import asyncio

# Use the HTTP crawler strategy
http_crawler_config = HTTPCrawlerConfig(
        method="GET",
        headers={"User-Agent": "MyCustomBot/1.0"},
        follow_redirects=True,
        verify_ssl=True
)

async def main():
    async with AsyncWebCrawler(crawler_strategy=AsyncHTTPCrawlerStrategy(browser_config =http_crawler_config)) as crawler:
        result = await crawler.arun("https://example.com")
        print(f"Status code: {result.status_code}")
        print(f"Content length: {len(result.html)}")

asyncio.run(main())
```

### 4. Docker Deployment

Crawl4AI can now be easily deployed as a Docker container, providing a
consistent and isolated environment. The Docker image includes a FastAPI server
with both streaming and non-streaming endpoints.

```bash
# Build the image (from the project root)
docker build -t crawl4ai .

# Run the container
docker run -d -p 8000:8000 --name crawl4ai crawl4ai
```

**API Endpoints:**

- `/crawl` (POST): Non-streaming crawl.
- `/crawl/stream` (POST): Streaming crawl (NDJSON).
- `/health` (GET): Health check.
- `/schema` (GET): Returns configuration schemas.
- `/md/{url}` (GET): Returns markdown content of the URL.
- `/llm/{url}` (GET): Returns LLM extracted content.
- `/token` (POST): Get JWT token

**Breaking Changes:**

- Docker deployment now requires a `.llm.env` file for API keys.
- Docker deployment now requires Redis and a new `config.yml` structure.
- Server startup now uses `supervisord` instead of direct process management.
- Docker server now requires authentication by default (JWT tokens).

See the [Docker deployment documentation](../../core/docker-deployment.md) for
detailed instructions.

### 5. Command-Line Interface (CLI)

A new CLI (`crwl`) provides convenient access to Crawl4AI's functionality from
the terminal.

```bash
# Basic crawl
crwl https://example.com

# Get markdown output
crwl https://example.com -o markdown

# Use a configuration file
crwl https://example.com -B browser.yml -C crawler.yml

# Use LLM-based extraction
crwl https://example.com -e extract.yml -s schema.json

# Ask a question about the crawled content
crwl https://example.com -q "What is the main topic?"

# See usage examples
crwl --example
```

See the [CLI documentation](../docs/md_v2/core/cli.md) for more details.

### 6. LXML Scraping Mode

Added `LXMLWebScrapingStrategy` for faster HTML parsing using the `lxml`
library. This can significantly improve scraping performance, especially for
large or complex pages. Set `scraping_strategy=LXMLWebScrapingStrategy()` in
your `CrawlerRunConfig`.

**Breaking Change:** The `ScrapingMode` enum has been replaced with a strategy
pattern. Use `WebScrapingStrategy` (default) or `LXMLWebScrapingStrategy`.

### 7. Proxy Rotation

Added `ProxyRotationStrategy` abstract base class with `RoundRobinProxyStrategy`
concrete implementation.

```python
import re
from crawl4ai import (
    AsyncWebCrawler,
    BrowserConfig,
    CrawlerRunConfig,
    CacheMode,
    RoundRobinProxyStrategy,
)
import asyncio
from crawl4ai.configs import ProxyConfig
async def main():
    # 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())
```

## Other Changes and Improvements

- **Added: `LLMContentFilter` for intelligent markdown generation.** This new
  filter uses an LLM to create more focused and relevant markdown output.

```python
from crawl4ai import AsyncWebCrawler, CrawlerRunConfig, DefaultMarkdownGenerator
from crawl4ai.content_filter_strategy import LLMContentFilter
from crawl4ai.async_configs import LlmConfig
import asyncio

llm_config = LlmConfig(provider="gemini/gemini-1.5-pro", api_token="env:GEMINI_API_KEY")

markdown_generator = DefaultMarkdownGenerator(
    content_filter=LLMContentFilter(llmConfig=llm_config, instruction="Extract key concepts and summaries")
)

config = CrawlerRunConfig(markdown_generator=markdown_generator)
async def main():
    async with AsyncWebCrawler() as crawler:
        result = await crawler.arun("https://docs.crawl4ai.com", config=config)
        print(result.markdown.fit_markdown)

asyncio.run(main())
```

- **Added: URL redirection tracking.** The crawler now automatically follows
  HTTP redirects (301, 302, 307, 308) and records the final URL in the
  `redirected_url` field of the `CrawlResult` object. No code changes are
  required to enable this; it's automatic.

- **Added: LLM-powered schema generation utility.** A new `generate_schema`
  method has been added to `JsonCssExtractionStrategy` and
  `JsonXPathExtractionStrategy`. This greatly simplifies creating extraction
  schemas.

```python
from crawl4ai.extraction_strategy import JsonCssExtractionStrategy
from crawl4ai.async_configs import LlmConfig

llm_config = LlmConfig(provider="gemini/gemini-1.5-pro", api_token="env:GEMINI_API_KEY")

schema = JsonCssExtractionStrategy.generate_schema(
    html="<div class='product'><h2>Product Name</h2><span class='price'>$99</span></div>",
    llmConfig = llm_config,
    query="Extract product name and price"
)
print(schema)
```

 Expected Output (may vary slightly due to LLM)
```JSON
{
  "name": "ProductExtractor",
  "baseSelector": "div.product",
  "fields": [
      {"name": "name", "selector": "h2", "type": "text"},
      {"name": "price", "selector": ".price", "type": "text"}
    ]
 }
```

- **Added: robots.txt compliance support.** The crawler can now respect
  `robots.txt` rules. Enable this by setting `check_robots_txt=True` in
  `CrawlerRunConfig`.

  ```python
  config = CrawlerRunConfig(check_robots_txt=True)
  ```

- **Added: PDF processing capabilities.** Crawl4AI can now extract text, images,
  and metadata from PDF files (both local and remote). This uses a new
  `PDFCrawlerStrategy` and `PDFContentScrapingStrategy`.

  ```python
  from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
  from crawl4ai.processors.pdf import PDFCrawlerStrategy, PDFContentScrapingStrategy
  import asyncio

  async def main():
      async with AsyncWebCrawler(crawler_strategy=PDFCrawlerStrategy()) as crawler:
          result = await crawler.arun(
              "https://arxiv.org/pdf/2310.06825.pdf",
              config=CrawlerRunConfig(
                  scraping_strategy=PDFContentScrapingStrategy()
              )
          )
          print(result.markdown)  # Access extracted text
          print(result.metadata)  # Access PDF metadata (title, author, etc.)

  asyncio.run(main())
  ```

- **Added: Support for frozenset serialization.** Improves configuration
  serialization, especially for sets of allowed/blocked domains. No code changes
  required.

- **Added: New `LlmConfig` parameter.** This new parameter can be passed for
  extraction, filtering, and schema generation tasks. It simplifies passing
  provider strings, API tokens, and base URLs across all sections where LLM
  configuration is necessary. It also enables reuse and allows for quick
  experimentation between different LLM configurations.

  ```python
  from crawl4ai.async_configs import LlmConfig
  from crawl4ai.extraction_strategy import LLMExtractionStrategy
  from crawl4ai import AsyncWebCrawler, CrawlerRunConfig

  # Example of using LlmConfig with LLMExtractionStrategy
  llm_config = LlmConfig(provider="openai/gpt-4o", api_token="YOUR_API_KEY")
  strategy = LLMExtractionStrategy(llmConfig=llm_config, schema=...)

  # Example usage within a crawler
  async with AsyncWebCrawler() as crawler:
      result = await crawler.arun(
          url="https://example.com",
          config=CrawlerRunConfig(extraction_strategy=strategy)
      )
  ```
  **Breaking Change:** Removed old parameters like `provider`, `api_token`,
  `base_url`, and `api_base` from `LLMExtractionStrategy` and
  `LLMContentFilter`. Users should migrate to using the `LlmConfig` object.

- **Changed: Improved browser context management and added shared data support.
  (Breaking Change:** `BrowserContext` API updated). Browser contexts are now
  managed more efficiently, reducing resource usage. A new `shared_data`
  dictionary is available in the `BrowserContext` to allow passing data between
  different stages of the crawling process. **Breaking Change:** The
  `BrowserContext` API has changed, and the old `get_context` method is
  deprecated.

- **Changed:** Renamed `final_url` to `redirected_url` in `CrawledURL`. This
  improves consistency and clarity. Update any code referencing the old field
  name.

- **Changed:** Improved type hints and removed unused files. This is an internal
  improvement and should not require code changes.

- **Changed:** Reorganized deep crawling functionality into dedicated module.
  (**Breaking Change:** Import paths for `DeepCrawlStrategy` and related classes
  have changed). This improves code organization. Update imports to use the new
  `crawl4ai.deep_crawling` module.

- **Changed:** Improved HTML handling and cleanup codebase. (**Breaking
  Change:** Removed `ssl_certificate.json` file). This removes an unused file.
  If you were relying on this file for custom certificate validation, you'll
  need to implement an alternative approach.

- **Changed:** Enhanced serialization and config handling. (**Breaking Change:**
  `FastFilterChain` has been replaced with `FilterChain`). This change
  simplifies config and improves the serialization.

- **Added:** Modified the license to Apache 2.0 _with a required attribution
  clause_. See the `LICENSE` file for details. All users must now clearly
  attribute the Crawl4AI project when using, distributing, or creating
  derivative works.

- **Fixed:** Prevent memory leaks by ensuring proper closure of Playwright
  pages. No code changes required.

- **Fixed:** Make model fields optional with default values (**Breaking
  Change:** Code relying on all fields being present may need adjustment).
  Fields in data models (like `CrawledURL`) are now optional, with default
  values (usually `None`). Update code to handle potential `None` values.

- **Fixed:** Adjust memory threshold and fix dispatcher initialization. This is
  an internal bug fix; no code changes are required.

- **Fixed:** Ensure proper exit after running doctor command. No code changes
  are required.
- **Fixed:** JsonCss selector and crawler improvements.
- **Fixed:** Not working long page screenshot (#403)
- **Documentation:** Updated documentation URLs to the new domain.
- **Documentation:** Added SERP API project example.
- **Documentation:** Added clarifying comments for CSS selector behavior.
- **Documentation:** Add Code of Conduct for the project (#410)

## Breaking Changes Summary

- **Dispatcher:** The `MemoryAdaptiveDispatcher` is now the default for
  `arun_many()`, changing concurrency behavior. The return type of `arun_many`
  depends on the `stream` parameter.
- **Deep Crawling:** `max_depth` is now part of `CrawlerRunConfig` and controls
  crawl depth. Import paths for deep crawling strategies have changed.
- **Browser Context:** The `BrowserContext` API has been updated.
- **Models:** Many fields in data models are now optional, with default values.
- **Scraping Mode:** `ScrapingMode` enum replaced by strategy pattern
  (`WebScrapingStrategy`, `LXMLWebScrapingStrategy`).
- **Content Filter:** Removed `content_filter` parameter from
  `CrawlerRunConfig`. Use extraction strategies or markdown generators with
  filters instead.
- **Removed:** Synchronous `WebCrawler`, CLI, and docs management functionality.
- **Docker:** Significant changes to Docker deployment, including new
  requirements and configuration.
- **File Removed**: Removed ssl_certificate.json file which might affect
  existing certificate validations
- **Renamed**: final_url to redirected_url for consistency
- **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

## Migration Guide

1. **Update Imports:** Adjust imports for `DeepCrawlStrategy`,
   `BreadthFirstSearchStrategy`, and related classes due to the new
   `deep_crawling` module structure.
2. **`CrawlerRunConfig`:** Move `max_depth` to `CrawlerRunConfig`. If using
   `content_filter`, migrate to an extraction strategy or a markdown generator
   with a filter.
3. **`arun_many()`:** Adapt code to the new `MemoryAdaptiveDispatcher` behavior
   and the return type.
4. **`BrowserContext`:** Update code using the `BrowserContext` API.
5. **Models:** Handle potential `None` values for optional fields in data
   models.
6. **Scraping:** Replace `ScrapingMode` enum with `WebScrapingStrategy` or
   `LXMLWebScrapingStrategy`.
7. **Docker:** Review the updated Docker documentation and adjust your
   deployment accordingly.
8. **CLI:** Migrate to the new `crwl` command and update any scripts using the
   old CLI.
9. **Proxy:**: Removed synchronous WebCrawler support and related rate limiting
   configurations.
10. **Config:**: Replace FastFilterChain to FilterChain