ai-hedge-fund

This is a code-simulated AI investment team. It uses Large Language Models (LLMs) to role-play 19 real-world investment masters like Buffett and Soros, analyzing financial reports and data like humans, then simulating meetings to give stock buy/sell recommendations.

It solves the problem where ordinary people can't understand complex quantitative trading logic. You don't need to actually trade to experience how AI analyzes market sentiment or calculates valuation. It turns boring financial analysis into a visualized AI conversation process, replacing traditional strategy research that requires deep financial knowledge.

1. Programmers interested in AI Agent (intelligent agent) development.

2. Investors who want to understand quantitative trading logic without capital for live trading.

3. Finance students wanting to observe decision-making differences between investment schools (e.g., value vs. growth).

4. Tech enthusiasts who want to play a "AI stock trading" simulation game.

1. Run locally in the terminal, input a stock ticker, and watch AI masters "debate" and reach a conclusion.

2. Learn multi-agent collaboration architecture and reference its code structure to build your own AI apps.

3. Observe AI reactions to breaking financial news to verify if the AI has risk awareness.

4. Use as a teaching case to show students the potential of LLMs in vertical fields like finance.

Compared to simple chatbots, this project builds a complete "Research-Risk Control-Decision" workflow. The 19 roles have clear divisions of labor, covering both fundamental and technical analysis, plus a dedicated risk manager. The architecture is clear and open-source, perfect for learning AI workflow design.

Requires some programming knowledge and Python environment setup. You must apply for and configure API Keys (e.g., OpenAI, DeepSeek), as it calls LLM interfaces, which may incur small fees.