TradingAgents
A research framework that coordinates analyst, researcher, and trader roles over market information. Its generated decisions are research artifacts, not investment advice.
Utility: Domain-specific; kind: framework.
Stars: 118k+; Starred in September: 7k+.
Metrics updated: 31 Oct 2026. Profile reviewed: 31 Oct 2026.
Utilities
- Study multi-agent financial analysis and debate workflows.
- Generate inspectable reports using configured market-data and model providers.
Who it is for
- Researchers evaluating agent-based financial-analysis methods.
Requirements
- Python 3.11+ or the documented Docker deployment.
- Configure model access and the market-data providers required by the chosen workflow.
Limitations
- The project explicitly disclaims financial, investment, and trading advice.
- Backtests and generated decisions depend on data timing, quality, assumptions, and nondeterministic model output.
Cost and privacy notes
- Data access, models, and repeated analysis/backtests can incur costs.
- Do not expose account credentials or treat generated decisions as authorization to trade.
Illustrative example (untested)
Illustrative, not tested: analyze a historical ticker/date and inspect whether every cited input was available at that time.
License
Apache-2.0; reviewed the repository LICENSE.