GitHub radar
TradingAgents: Multi-Agent AI Stock Analysis
107k-star multi-agent system from TauricResearch that mirrors a trading firm — specialist AI agents research fundamentals, sentiment, news, and technicals, then debate and reach a decision together.
TradingAgents is a multi-agent stock analysis tool developed by TauricResearch. It aims to replicate the organization of a real-world trading company. The Fundamentals Analyst examines financial statements and earnings reports. The Sentiment Analyst gathers information from various sources like StockTwits, Reddit, and news articles to get an overall sentiment of the market. The News Analyst keeps track of global news and its effect on stock prices. The Technical Analyst looks at charts and technical indicators like MACD and RSI. Finally, two researchers (one bullish, the other bearish) engage in a formal debate based on the findings of the previous agents. The risk management team assesses risk in the portfolio while the Trader Agent makes the ultimate call. All of these agents operate within a single CLI interface where you can input the stock ticker, analysis date, and desired LLM engine, and kick off the entire multi-agent workflow. With version 0.5.0, we’ve introduced backtesting across a range of tickers and dates, as well as access to SEC EDGAR fundamentals data (e.g., when running a backtest prior to Apple’s restatement of their 2010 filing, the system will use the 2008 data instead). We support any market that Yahoo Finance supports (US, HK, London, Tokyo, India, etc.).
Why a vibe-coder should care
This is a great way to get a feel for how multi-agent debates work, especially when you consider the various perspectives, formalized disagreement, and final decision making all happening at once! If you’re interested in the stock market, this is also a nice structured way to conduct multi-perspective research on a particular stock without having to manually collect data yourself. Remember, this is just a research tool, not a trading bot!
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Deploy TradingAgents from https://github.com/TauricResearch/TradingAgents — run cp .env.example .env, ask me for ANTHROPIC_API_KEY and ALPHA_VANTAGE_API_KEY, start docker compose run --rm tradingagents and show me an analysis of AAPL.
Runs on a regular laptop — needs an API key from an LLM provider (Claude, OpenAI, or others) and a free Alpha Vantage key for market data. Simplest path is Docker: one command after setting up the .env file.
Open on GitHub▌ More finds