TauricResearch/TradingAgents
TradingAgents: Multi-Agents LLM Financial Trading Framework
TradingAgents is an open-source Python framework simulating a professional trading team with multi-agent AI. Specialized agents cover fundamentals, technical analysis, news sentiment, and risk management, then coordinate to produce structured buy/hold/sell recommendations.
Our Review
Finance is one of the domains where multi-agent architecture maps most directly to existing professional practice. Trading desks have always used specialist teams -- macro analyst, technicals analyst, risk manager -- where the final position reflects a synthesis of expert views rather than a single opinion. TauricResearch formalized this structure for LLMs in TradingAgents: each specialist role is a separate agent with its own data scope and reasoning chain, and a portfolio manager agent synthesizes all reports into a structured final recommendation.
Key capabilities
- •Multi-agent analyst team: separate agents analyze earnings fundamentals, price technicals (RSI, MACD, moving averages), news sentiment, and macro context independently before synthesis
- •Portfolio manager synthesis: a coordinator agent aggregates all analyst outputs and produces a structured buy/hold/sell recommendation with reasoning and confidence scoring
- •Multi-LLM backend support: any OpenAI-compatible model serves as the analyst backbone -- Claude, GPT-4o, DeepSeek, Gemini, or locally-hosted models via Ollama
- •Configurable team composition: add or remove analyst roles per use case; a quick screen needs fewer agents than a deep research report
- •Historical backtesting: run the agent team against historical periods to evaluate decision quality before deploying in live analysis workflows
- •Structured JSON output: each agent produces typed reports that integrate with downstream dashboards, alerts, or automated research pipelines
Getting started
Clone TauricResearch/TradingAgents and install dependencies with pip install -r requirements.txt. Set your LLM API key and financial data provider credentials (Polygon.io, Yahoo Finance, or Finnhub). Run python main.py with a ticker symbol to invoke the full analyst team and receive a structured research report.
Limitation
TradingAgents produces analysis and signals, not live trade execution -- integrating with a brokerage API for automated trading requires additional engineering. Real-time financial data with full coverage requires paid subscriptions (Polygon.io from $29/month). Smaller LLMs produce noticeably weaker analyst reasoning; GPT-4o or Claude Sonnet-class models are recommended for production analysis. LLM-generated trading signals should always be reviewed by a qualified person before any financial decision.
Our Verdict
TradingAgents applies a structurally sound idea to financial analysis in 2026 -- mirroring how professional trading desks work by assigning each type of analysis to a dedicated agent with a specific role and data scope. The result is more structured than asking a single model for a stock opinion: fundamentals, technicals, and sentiment each get their own reasoning chain before synthesis.
The 76,000 GitHub stars reflect genuine developer interest in LLM-powered financial workflows. For teams building AI-assisted investment research tools, TradingAgents is a functional foundation: the analyst team is configurable, the output is structured JSON, and any OpenAI-compatible model can serve as the backend without framework changes.
The realistic framing is that LLM-based trading analysis is a research aid, not a trading system. The framework produces human-readable reports that investors can interpret and validate; it does not generate alpha directly or handle trade execution. Teams evaluating TradingAgents should focus on the workflow architecture and validate signal quality against historical periods carefully before using outputs to inform real financial decisions in 2026.
Frequently Asked Questions
What is TradingAgents and how does it work?
TradingAgents simulates a professional trading team using multi-agent AI. Each agent -- fundamentals analyst, technical analyst, news sentiment analyst, and risk manager -- independently analyzes a stock from its domain, then a portfolio manager agent synthesizes all reports into a structured buy/hold/sell recommendation in 2026.
Which AI models does TradingAgents support?
TradingAgents supports any OpenAI-compatible LLM as the analyst backend. This includes OpenAI GPT-4o, Anthropic Claude, Google Gemini, DeepSeek, Qwen, and locally-hosted models via Ollama or LM Studio. Each agent in the team can use a different model, though consistent model tiers produce better coordinated analysis.
Can TradingAgents execute trades automatically?
No. TradingAgents is a research and analysis framework, not a trading bot. It produces structured analyst reports and buy/hold/sell recommendations, but does not connect to brokerage APIs or execute orders. Integrating automated execution requires building a separate trading layer on top of TradingAgents' structured output signals.
What financial data sources does TradingAgents support?
TradingAgents integrates with Polygon.io, Yahoo Finance, Finnhub, and other market data providers for price history, fundamentals, and news feeds. Yahoo Finance works on the free tier but has rate limits. Real-time data with full coverage requires a paid Polygon.io or Finnhub subscription. The data layer is configurable for additional providers.
How does TradingAgents compare to a rules-based trading algorithm?
Rules-based algorithms follow deterministic logic -- buy when RSI crosses 30, sell on MACD divergence. TradingAgents uses LLM reasoning to analyze qualitative factors like news sentiment, earnings call tone, and analyst commentary alongside quantitative signals, then synthesizes them as a structured opinion. It complements rules-based systems rather than replacing them.
What is TradingAgents?
TradingAgents is an open-source Python framework simulating a professional trading team with multi-agent AI. Specialized agents cover fundamentals, technical analysis, news sentiment, and risk management, then coordinate to produce structured buy/hold/sell recommendations.
How do I install TradingAgents?
Visit the GitHub repository at https://github.com/TauricResearch/TradingAgents for installation instructions.
What license does TradingAgents use?
TradingAgents uses the Apache-2.0 license.
What are alternatives to TradingAgents?
Explore related tools and alternatives on My AI Guide.
Open source & community-verified
Apache-2.0 licensed: free to use in any project, no strings attached. 77,722 developers have starred this, meaning the community has reviewed and trusted it.
Reviewed by My AI Guide for relevance, quality, and active maintenance before listing.
Topics