AI · GitHub repo
ai-hedge-fund
Proof-of-concept AI multi-agent trading system with paper trading and backtesting.
ai-hedge-fund is an open-source terminal application that uses multiple AI agents to simulate investment decisions inspired by famous investor methodologies like Warren Buffett. It gathers market fundamentals, technicals, and earnings data, synthesizes them through risk and portfolio management modules, and outputs actionable signals. Designed for education and research, it features paper trading and backtesting modes without executing live trades.
What it's for
- Simulate stock trading strategies using specialized AI analyst personas
- Backtest investment ideas against historical benchmark data
- Paper trade simulated portfolios using a local hash-chained ledger
- Study multi-agent decision architecture applied to financial markets
Features
- Interactive terminal user interface (TUI) installed via pipx or uv
- Multiple agent personas covering fundamentals, valuation, technicals, and Warren Buffett-style value investing
- Portfolio and risk management layers that generate unified buy, hold, or sell decisions with confidence scores
- Paper trading mode with ledger tracking and session history
- Historical backtesting mode with equity curve visualizations against benchmarks
- Support for multiple LLM providers including OpenAI, Anthropic, DeepSeek, Google, and xAI
- Integration with Financial Datasets API for market metrics and pricing
Pricing
Open source. Free and open source (MIT); requires user-provided API keys for LLMs and Financial Datasets
Details
- Type
- GitHub repo
- Category
- AI
- Topic
- Agent Frameworks
- Open source
- Yes
- Platforms
- Command line, Linux, macOS, Windows
- For
- Developers, quantitative finance students, and algorithmic trading enthusiasts exploring multi-agent LLM systems.
- Tags
- multi agent, trading, finance, cli, python, backtesting, paper trading
Why it was saved
- @virattt: A working multi-agent trading system: pick analyst agents in a terminal UI (Technical, Fundamentals, Sentiment, Valuation, Warren Buffett), the agents analyze tickers, a Risk Management and Portfolio Management step produce trading decisions with confidence scores and reasoning.
Related
Lightweight open-source Python library for building code-writing and vision AI agents
Open-source multi-agent framework for building autonomous social and on-chain AI agents
Open-source persistent memory layer for autonomous AI agents using knowledge graphs
A workflow pattern where agents document completed tasks into reusable shortcuts
AI agent skill for deep social research across Reddit, X, and more.
Multi-agent system querying multiple LLMs that peer-review answers into one synthesis.
Updated