Open-Source AI Framework Simulates Trading Firms With Multi-Agent LLM System, Surpasses 55,000 GitHub Stars

TradingAgents, an open-source multi-agent LLM framework that replicates real-world trading firms using specialized AI roles, has surpassed 55,000 GitHub stars as its latest version introduces crash recovery, persistent decision logging, and expanded support for major AI providers including OpenAI, Google, and DeepSeek.

• TradingAgents is an open-source multi-agent LLM framework that simulates real-world trading firms by deploying specialized AI agents — including fundamental, sentiment, news, and technical analysts — alongside trader, risk management, and portfolio manager roles to collaboratively make financial trading decisions.
• The framework supports multiple LLM providers including OpenAI, Google, Anthropic, xAI, DeepSeek, Qwen, GLM, OpenRouter, Ollama, and Azure OpenAI, and can be installed via pip or Docker with a CLI interface for selecting tickers, analysis dates, and research depth.
• Version 0.2.4 introduces structured-output agents, LangGraph checkpoint resume for crash recovery, a persistent decision log that tracks past trade performance and injects lessons into future analyses, and expanded provider support — with the project now boasting over 55,000 GitHub stars and 10,000 forks.