Multi-Agent Quantitative Trading System

An advanced algorithmic trading simulator utilizing a multi-agent AI architecture and classical quantitative indicators.

Architecture

• Data Layer: Fetches and standardizes historical OHLCV data via Yahoo Finance.
• AI Agents (agents/):
  • Analyst: Generates text summarizing multi-day price-action trends.
  • Strategist: Processes quantitative inputs to output directional conviction.
  • Risk Manager: Emits strict position sizing rules and stop-loss bounds.
  • Trader: Final executive decision-maker parsing cross-agent signals.
• Strategy Engine (strategies/): Technical indicators (RSI, SMA, MACD) and momentum/mean-reversion baseline logic.
• Execution & Backtesting (engine/): Virtual portfolio tracking cash, PnL, margin, and standard metrics (Sharpe Ratio, Max Drawdown).

Setup & Execution

1. Create and activate a Virtual Environment.
2. Install dependencies: pip install -r requirements.txt.
3. Provide an OpenAI API Key in .env or in the UI.
4. Run the Streamlit Application:

   ./run.sh
   # or
   streamlit run app.py