TradingAgents Multi-Agent LLM Trading Framework
TL;DR
TradingAgents is an open-source multi-agent trading framework that mirrors real-world trading firm dynamics, deploying specialized LLM-powered agents for fundamental analysis, sentiment, news, technical analysis, risk management, and portfolio management to collaboratively generate trading decisions. It decomposes complex trading tasks across multiple specialized AI agents that debate bullish and bearish cases, producing more structured and informed trading insights than a single AI model — enabling developers to build sophisticated fintech tools or research workflows.
What it actually is
- What: TradingAgents is an open-source multi-agent trading framework that mirrors real-world trading firm dynamics, deploying specialized LLM-powered agents for fundamental analysis, sentiment, news, technical analysis, risk management, and portfolio management to collaboratively generate trading decisions.
- Who built it / maintained by: Tauric Research
- Status: stable
- Why it matters: It decomposes complex trading tasks across multiple specialized AI agents that debate bullish and bearish cases, producing more structured and informed trading insights than a single AI model — enabling developers to build sophisticated fintech tools or research workflows.
- How it compares to alternatives:
- FinRobot
- FinAgent
- OpenBB
- Composer
- Alpaca AI
- AutoGen finance workflows
- GitHub stars: 87,813 · License: Apache-2.0 · Archived: no
Links
Kickstarter guide
Clone the repository from github.com/TauricResearch/TradingAgents and install dependencies via pip. Configure your preferred LLM backbone (supports DeepSeek, Qwen, GLM, Azure, and others) and set up API keys for market data. Run the framework using Docker or directly on your machine (Windows UTF-8 support included). Review the README for agent configuration and the LangGraph checkpoint system for persistent decision logging.