An open-source, research-driven multi-agent trading framework powered by LangGraph and LLMs. Seven specialized AI agents analyze, debate, and execute trading strategies together.
From multi-model support to 10+ integrated data sources — TradingAgents is built for serious research and real-world trading.
OpenAI, DeepSeek, Claude, Qwen, Llama, Gemini — assign different models to different agent roles based on their strengths.
Structured multi-round debate between agents produces more robust trading decisions through rigorous analysis.
Full audit trail of every decision, debate, and trade — transparent and reproducible for compliance and research.
Swap components without rewriting. Plug-and-play design for rapid experimentation with new strategies.
Tushare, Yahoo Finance, Alpha Vantage, Wind, and more — global markets covered through a unified interface.
Grounded in academic research with published papers and experimental results. Built for researchers, by researchers.
Hear from the community building the future of multi-agent trading.
"The multi-agent debate system is unlike anything I've seen. Each agent brings a different perspective — it's like having a team of analysts working 24/7."
"Backtesting has never been this thorough. The transaction logs alone are worth it — every decision is recorded and auditable. Perfect for our research team."
"The modular architecture is a game-changer. We swapped out the data source and model provider in under an hour. Incredible flexibility for experimentation."
From open-source to hosted platform to private deployment — find the right fit for your goals.
Connect to 10+ financial data providers and all major LLM platforms — all managed through a unified config interface.
Share your thoughts, questions, or feedback about TradingAgents.
Your message will be displayed publicly after submission.