TradingAgents
Multi-agent LLM framework that simulates a trading firm's research desk
What it is
A Python framework where specialized LLM agents act as fundamentals, sentiment, news, and technical analysts, then debate as researchers before a trader and risk team settle on a decision for a ticker. It pulls market data from sources such as Yahoo, SEC EDGAR, and FRED, supports many model providers, and includes a CLI, backtesting over ticker and date grids, and HTML reports. Researchers studying how agent teams reason about markets can run it from the command line or import it as a package; it is not presented as investment advice.