📈AI Hedge Fund Open Source Com 18 Agentes Inspirados em Warren Buffett
Alguém abriu o código de um fundo de investimento totalmente gerido por IA — com 18 agentes especializados, cada um programado para pensar como um grande investidor: Warren Buffett, Charlie Munger, George Soros, Cathie Wood. Cada agente analisa uma ação de um ângulo diferente, e o sistema consolida as opiniões. --- É um experimento, não um fundo real. Mas a ideia é fascinante: em vez de um modelo de IA genérico tentando fazer tudo, você tem uma mesa de especialistas — cada um com sua filosofia, seu viés, seu estilo de análise. É assim que boas decisões de investimento são tomadas entre humanos. A questão é se dá pra replicar isso em código.
🚨 Someone just open sourced an AI hedge fund with 18 agents that think like Wall Street legends. Warren Buffett. Charlie Munger. Michael Burry. Cathie Wood. Bill Ackman. All running on your laptop. It's called AI Hedge Fund. You give it stock tickers. 18 AI agents analyze the company from every angle. Then they vote on whether to buy, sell, or hold. Not a toy. Not a dashboard. A full multi-agent investment research system. No Bloomberg Terminal. No $25K brokerage minimums. No financial advisor fees. Just AI agents doing what hedge funds charge 2-and-20 for. Here's who's on your team: → Warren Buffett Agent. Only buys wonderful businesses at fair prices → Charlie Munger Agent. Demands a margin of safety on every pick → Michael Burry Agent. The Big Short contrarian hunting deep value → Cathie Wood Agent. Innovation and disruption. High conviction growth → Bill Ackman Agent. Activist investor. Takes bold positions → Ben Graham Agent. The godfather of value investing. Hidden gems only → Aswath Damodaran Agent. The Dean of Valuation. Story meets numbers → Plus 11 more specialized agents covering technicals, sentiment, risk, and fundamentals Here's how it works: → You enter stock tickers (AAPL, NVDA, TSLA, whatever you want) → Agents pull real financial data. Earnings, balance sheets, insider trades, news → Each agent analyzes the data through their own investment philosophy → A Risk Manager agent checks position sizing and portfolio exposure → A Portfolio Manager agent takes all signals and makes the final call → You get a buy/sell/hold decision with full reasoning from every agent Here's the wildest part: You can turn on --show-reasoning and watch each agent explain their logic step by step. Warren Buffett agent breaks down the moat. Michael Burry agent flags the hidden risks. Cathie Wood agent finds the disruption angle. They literally argue with each other. It has a full backtester. Run your strategy against historical data and see how it would have performed. Full web UI included. Not just a terminal tool. A real dashboard. Works with OpenAI, Claude, Groq, DeepSeek, or fully local with Ollama. Your data never has to leave your machine. Data for AAPL, GOOGL, MSFT, NVDA, and TSLA is completely free. No API key needed. 46.7K GitHub stars. 8.1K forks. Actively maintained. 100% Open Source. MIT License.
— @heynavtoor View on X
