This project provides a Chinese-language enhanced version of the TradingAgents framework, designed for financial trading decisions using multi-agent LLMs. It targets Chinese users and developers seeking localized support, A-share market integration, and domestic LLM compatibility, offering a user-friendly web interface and comprehensive documentation.

How It Works

The framework leverages a multi-agent architecture, simulating a trading company's structure with specialized agents (analysts, researchers, traders, risk managers) collaborating to make trading decisions. It integrates various LLMs, including domestic Chinese models like Alibaba's Qwen series and Google's Gemini, alongside international options. The system emphasizes robust data handling through multi-layer caching (Redis, MongoDB) and fallback mechanisms to ensure reliability and performance.

Quick Start & Requirements

• Install: Clone the repository, create a virtual environment, and install dependencies using pip install -r requirements.txt. Optional database support requires pip install -r requirements_db.txt.
• Prerequisites: Python 3.10+ (3.11 recommended), 4GB+ RAM (8GB+ recommended).
• Configuration: Requires API keys for LLMs and data sources (e.g., DashScope, Finnhub). Database setup (MongoDB, Redis) is optional but recommended for performance.
• Documentation: Comprehensive Chinese documentation is available in the docs/ directory.

Highlighted Details

• Fully functional Web UI built with Streamlit for configuration and monitoring.
• Integrated A-share market data support via Tongdaxin API.
• Support for domestic Chinese LLMs (Alibaba Qwen) and international models (Gemini, OpenAI, Anthropic).
• Multi-layer data caching with Redis and MongoDB for performance and reliability.
• Detailed Chinese documentation exceeding 50,000 words, including architecture, examples, and troubleshooting.

Maintenance & Community

The project is actively maintained, with version 0.1.4 released recently. Community interaction is primarily through GitHub Issues.

Licensing & Compatibility

The project is licensed under the Apache 2.0 license, permitting commercial use, modification, and distribution with attribution.

Limitations & Caveats

The framework is intended for research and educational purposes only and does not constitute investment advice. AI model predictions carry inherent uncertainties, and trading performance can vary significantly. Users should exercise caution and consult financial advisors for investment decisions.