Metadata-Version: 2.4
Name: ai-hedge-fund
Version: 0.1.1
Summary: AI hedge fund analysis as Claude Code slash commands. 30+ AI investor personas analyze stocks.
Author: Naman Vinayak
License: MIT
Project-URL: Homepage, https://github.com/NamanVinayak/ai-hedge-fund
Project-URL: Repository, https://github.com/NamanVinayak/ai-hedge-fund
Keywords: claude-code,skill,hedge-fund,stock-analysis,ai-investing
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Financial and Insurance Industry
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Office/Business :: Financial :: Investment
Requires-Python: >=3.10
Description-Content-Type: text/markdown
Requires-Dist: yfinance>=0.2.40
Requires-Dist: pandas>=2.1.0
Requires-Dist: numpy>=1.24.0
Requires-Dist: requests>=2.31.0
Requires-Dist: pydantic>=2.4.2
Requires-Dist: python-dotenv>=1.0.0
Requires-Dist: sqlalchemy>=2.0.22
Requires-Dist: colorama>=0.4.6
Requires-Dist: questionary>=2.0.0
Requires-Dist: finnhub-python>=2.4.20
Requires-Dist: langchain-core>=0.1.0
Provides-Extra: extras
Requires-Dist: pandas_ta; extra == "extras"

# AI Hedge Fund — Claude Code Skill Collection

Analyze stocks using 30+ AI agents — legendary investor personas, swing trade strategies, and intraday day-trade systems. Built as a collection of Claude Code slash commands.

## Install

```bash
pip install ai-hedge-fund
ai-hedge-fund install
```

## Commands

| Command | What it does | Agents | Time |
|---------|-------------|--------|------|
| `/invest AAPL,MSFT` | Long-term portfolio decisions | 14 investor personas (Buffett, Munger, Graham, etc.) | ~5 min |
| `/swing TSLA` | Swing trade setups (2-20 days) | 9 swing strategies + Head Trader | ~3 min |
| `/daytrade SPY` | Intraday trade plan | 9 day-trade strategies + Head Trader | ~3 min |
| `/research NVDA` | Comprehensive research report | All 30+ agents | ~15 min |

## How It Works

Each command follows the same pattern:

```
Your ticker → Fetch financial data → N AI agents analyze independently
→ Deterministic math agents → Synthesis → Final decision/report
```

**Invest**: 14 legendary investors (Warren Buffett, Charlie Munger, Ben Graham, Cathie Wood, Michael Burry, etc.) each independently analyze the stock using their real-world investment philosophy. A portfolio manager synthesizes all opinions into a buy/sell/hold decision with holding duration.

**Swing**: 9 trading strategies (trend following, pullback, breakout, momentum, mean reversion, catalyst, sector rotation) analyze the daily chart. A Head Trader resolves conflicting signals. Output includes entry price, target, stop-loss, and risk-reward ratio.

**Day Trade**: 9 intraday strategies (VWAP, momentum scalping, mean reversion, breakout, gap analysis, volume profiling, pattern reading, stat arb, news catalyst) analyze 5-minute charts. Output is a specific trade plan for the day.

**Research**: All 30+ agents from all modes run. Output is a balanced report: bull case, bear case, key metrics, risk factors, sentiment breakdown. No trading recommendation.

## Data Sources (all free)

- **yfinance** — daily + intraday price data, market cap
- **SEC EDGAR** — 10+ years of quarterly financials, TTM snapshots
- **Finnhub** (free tier) — insider trades, company news

No paid API keys required. Works out of the box with prices, financials, and technicals.

### Optional: Finnhub API key (free, recommended)

For insider trading data and company news sentiment, add a free Finnhub API key:

1. Go to [finnhub.io](https://finnhub.io) and create a free account (no credit card needed)
2. Copy your API key from the dashboard
3. Create a `.env` file in your project directory:
```
FINNHUB_API_KEY=your_key_here
```

Without this key, the system still works — you just won't get insider trade patterns or news sentiment analysis.

## Uninstall

```bash
ai-hedge-fund uninstall
pip uninstall ai-hedge-fund
```

## Requirements

- Python 3.10+
- Claude Code

## License

MIT
