Created ARCHITECTURE.md as a complete technical reference for the
Crawl4AI Docker server, replacing the stress test pipeline document
with production-grade documentation.
Contents:
- System overview with architecture diagrams
- Core components deep-dive (server, API, utils)
- Smart browser pool implementation details
- Real-time monitoring system architecture
- WebSocket implementation and fallback strategy
- Memory management and container detection
- Production optimizations and code review fixes
- Deployment guides (local, Docker, production)
- Comprehensive troubleshooting section
- Debug tools and performance tuning
- Test suite documentation
- Architecture decision log (ADRs)
Target audience: Developers maintaining or extending the system
Goal: Enable rapid onboarding and confident modifications
Backend Improvements (11 fixes applied):
Critical Fixes:
- Add lock protection for browser pool access in monitor stats
- Ensure async track_janitor_event across all call sites
- Improve error handling in monitor request tracking (already in place)
Important Fixes:
- Replace fire-and-forget Redis with background persistence worker
- Add time-based expiry for completed requests/errors (5min cleanup)
- Implement input validation for monitor route parameters
- Add 4s timeout to timeline updater to prevent hangs
- Add warning when killing browsers with active requests
- Implement monitor cleanup on shutdown with final persistence
- Document memory estimates with TODO for actual tracking
Frontend Enhancements:
WebSocket Real-time Updates:
- Add WebSocket endpoint at /monitor/ws for live monitoring
- Implement auto-reconnect with exponential backoff (max 5 attempts)
- Add graceful fallback to HTTP polling on WebSocket failure
- Send comprehensive updates every 2 seconds (health, requests, browsers, timeline, events)
UI/UX Improvements:
- Add live connection status indicator with pulsing animation
- Green "Live" = WebSocket connected
- Yellow "Connecting..." = Attempting connection
- Blue "Polling" = Fallback to HTTP polling
- Red "Disconnected" = Connection failed
- Restore original beautiful styling for all sections
- Improve request table layout with flex-grow for URL column
- Add browser type text labels alongside emojis
- Add flex layout to browser section header
Testing:
- Add test-websocket.py for WebSocket validation
- All 7 integration tests passing successfully
Summary: 563 additions across 6 files
- Introduced a demo script (`demo_monitor_dashboard.py`) to showcase various monitoring features through simulated activity.
- Implemented a test script (`test_monitor_demo.py`) to generate dashboard activity and verify monitor health and endpoint statistics.
- Added a logo image to the static assets for branding purposes.
The library no longer supports Python 3.9 and so it was important to drop all references to python 3.9.
Following changes have been made:
- pyproject.toml: set requires-python to ">=3.10"; remove 3.9 classifier
- setup.py: set python_requires to ">=3.10"; remove 3.9 classifier
- docs: update Python version mentions
- deploy/docker/c4ai-doc-context.md: options -> 3.10, 3.11, 3.12, 3.13
- Updated ProxyConfig.from_string to support multiple proxy formats, including URLs with credentials.
- Deprecated the 'proxy' parameter in BrowserConfig, replacing it with 'proxy_config' for better flexibility.
- Added warnings for deprecated usage and clarified behavior when both parameters are provided.
- Updated documentation and tests to reflect changes in proxy configuration handling.
- Return comprehensive error messages along with status codes for api internal errors.
- Fix fit_html property serialization issue in both /crawl and /crawl/stream endpoints
- Add sanitization to ensure fit_html is always JSON-serializable (string or None)
- Add comprehensive error handling test suite.
Implement hierarchical configuration for LLM parameters with support for:
- Temperature control (0.0-2.0) to adjust response creativity
- Custom base_url for proxy servers and alternative endpoints
- 4-tier priority: request params > provider env > global env > defaults
Add helper functions in utils.py, update API schemas and handlers,
support environment variables (LLM_TEMPERATURE, OPENAI_TEMPERATURE, etc.),
and provide comprehensive documentation with examples.
Previously, the system incorrectly used OPENAI_API_KEY for all LLM providers
due to a hardcoded api_key_env fallback in config.yml. This caused authentication
errors when using non-OpenAI providers like Gemini.
Changes:
- Remove api_key_env from config.yml to let litellm handle provider-specific env vars
- Simplify get_llm_api_key() to return None, allowing litellm to auto-detect keys
- Update validate_llm_provider() to trust litellm's built-in key detection
- Update documentation to reflect the new automatic key handling
The fix leverages litellm's existing capability to automatically find the correct
environment variable for each provider (OPENAI_API_KEY, GEMINI_API_TOKEN, etc.)
without manual configuration.
ref #1291
- Add raw HTML URL validation alongside http/https checks
- Fix URL preprocessing logic to handle raw: and raw:// prefixes
- Update error message and add comprehensive test cases
- Support LLM_PROVIDER env var to override default provider (openai/gpt-4o-mini)
- Add optional 'provider' parameter to API endpoints for per-request overrides
- Implement provider validation to ensure API keys exist
- Update documentation and examples with new configuration options
Closes the need to hardcode providers in config.yml
- Bump version to 0.7.0
- Add release notes and demo files
- Update README with v0.7.0 features
- Update Docker configurations for v0.7.0-r1
- Move v0.7.0 demo files to releases_review
- Fix BM25 scoring bug in URLSeeder
Major features:
- Adaptive Crawling with pattern learning
- Virtual Scroll support for infinite pages
- Link Preview with 3-layer scoring
- Async URL Seeder for massive discovery
- Performance optimizations
- Fixed widespread typo: `temprature` → `temperature` across LLMConfig and related files
- Enhanced CSS/XPath selector guidance for more reliable LinkedIn data extraction
- Added Google Colab display server support for running Crawl4AI in notebook environments
- Improved browser debugging with verbose startup args logging
- Updated LinkedIn schemas and HTML snippets for better parsing accuracy
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Revised the description for the `f` parameter in the `/mcp/md` tool schema to use lowercase enum values
(`raw`, `fit`, `bm25`, `llm`) for consistency with the actual `enum` definition. This change prevents
LLM-based clients (e.g., Gemini via LibreChat) from generating uppercase values like `"FIT"`, which
caused 422 validation errors due to strict case-sensitive matching.
Implements new asynchronous endpoints for handling long-running crawl and LLM tasks:
- POST /crawl/job and GET /crawl/job/{task_id} for crawl operations
- POST /llm/job and GET /llm/job/{task_id} for LLM operations
- Added Redis-based task management with configurable TTL
- Moved schema definitions to dedicated schemas.py
- Added example polling client demo_docker_polling.py
This change allows clients to handle long-running operations asynchronously through a polling pattern rather than holding connections open.
- Add tables field to CrawlResult model while maintaining backward compatibility
- Update async_webcrawler.py to extract tables from media and pass to tables field
- Update crypto_analysis_example.py to use the new tables field
- Add /config/dump examples to demo_docker_api.py
- Bump version to 0.6.1
- Update Docker base image to Python 3.12-slim-bookworm
- Bump version from 0.6.0rc1 to 0.6.0
- Update documentation to reflect release version changes
- Fix license specification in pyproject.toml and setup.py
- Clean up code formatting in demo_docker_api.py
BREAKING CHANGE: Base Python version upgraded from 3.10 to 3.12
Major updates to Docker deployment infrastructure:
- Switch default port to 11235 for all services
- Add MCP (Model Context Protocol) support with WebSocket/SSE endpoints
- Simplify docker-compose.yml with auto-platform detection
- Update documentation with new features and examples
- Consolidate configuration and improve resource management
BREAKING CHANGE: Default port changed from 8020 to 11235. Update your configurations and deployment scripts accordingly.
