TradingAgents, a multi-agent LLM framework for financial trading, is open-sourced with support for multiple language models and structured agent roles.

TradingAgents framework open-sourced with specialized LLM-powered agents including Fundamentals Analyst, Sentiment Analyst, News Analyst, Technical Analyst, Researcher Team, Trader Agent, Risk Management, and Portfolio Manager roles that collaborate to evaluate market conditions and inform trading decisions.

Supports multiple LLM providers: OpenAI (GPT), Google (Gemini), Anthropic (Claude), xAI (Grok), DeepSeek, Qwen (Alibaba DashScope), GLM (Zhipu), OpenRouter, Ollama for local models, and Azure OpenAI for enterprise; built with LangGraph for flexibility and modularity.

Latest release (v0.2.4, April 2026) includes structured-output agents, LangGraph checkpoint resume, persistent decision log that tracks realized returns and injects past decisions into future analysis, and support for DeepSeek, Qwen, GLM, and Azure providers.