This major update adds four powerful features to TradingAgents:
1. Multi-LLM Provider Support
- LLMFactory for OpenAI, Anthropic Claude, and Google Gemini
- Easy provider switching via configuration
- Recommended models for each provider
- Updated TradingAgentsGraph to use factory pattern
2. Paper Trading Integration
- BaseBroker abstract interface for consistency
- AlpacaBroker implementation with free paper trading
- Support for market, limit, stop, and stop-limit orders
- Real-time position tracking and P&L calculation
- Example scripts for basic and integrated trading
3. Web Interface
- Beautiful Chainlit-based GUI
- Chat interface for stock analysis
- Interactive trading commands
- Portfolio management
- Settings configuration
- Real-time updates
4. Docker Support
- Production-ready Dockerfile
- Docker Compose for multi-service setup
- Persistent data volumes
- Optional Jupyter notebook service
- Comprehensive deployment documentation
Additional improvements:
- Enhanced .env.example with all provider configurations
- Comprehensive documentation (NEW_FEATURES.md, DOCKER.md)
- Broker integration guide
- Example scripts for all features
- Verification script to test new features
- Made example scripts executable
Files changed:
- New: tradingagents/llm_factory.py (400+ lines)
- New: tradingagents/brokers/ (base.py, alpaca_broker.py, __init__.py)
- New: web_app.py (Chainlit web interface)
- New: Dockerfile, docker-compose.yml, .dockerignore
- New: examples/use_claude.py, paper_trading_alpaca.py, tradingagents_with_alpaca.py
- New: NEW_FEATURES.md, DOCKER.md, tradingagents/brokers/README.md
- New: verify_new_features.py
- Modified: tradingagents/graph/trading_graph.py (use LLMFactory)
- Modified: .env.example (added all providers)
All features verified and tested.
- Added support for running CLI and Ollama server via Docker
- Introduced tests for local embeddings model and standalone Docker setup
- Enabled conditional Ollama server launch via LLM_PROVIDER
Claude
Yijia Xiao
Geeta Chauhan