This repository details a 6-month live trading experiment using ChatGPT to manage a real-money micro-cap stock portfolio, starting with $100. It aims to determine if LLMs can generate alpha or make intelligent trading decisions with real-time data, targeting individuals interested in AI's practical application in finance and algorithmic trading.

How It Works

The system provides ChatGPT with daily trading data for its portfolio and allows weekly deep research evaluations. Strict stop-loss rules are enforced. The core decision-making is handled by ChatGPT, with Python scripts managing price evaluation, holding updates, and performance tracking. This approach leverages LLM capabilities for potentially novel trading strategies, with all data and decisions logged for transparency.

Quick Start & Requirements

• Install dependencies: pip install -r requirements.txt
• Requires Python and libraries listed in requirements.txt (e.g., Pandas, yFinance, Matplotlib).
• Access to ChatGPT API is implicitly required for the decision-making engine.
• The experiment runs from June 2025 to December 2025, with daily portfolio updates.

Highlighted Details

• Live trading scripts for daily evaluation and holding updates.
• LLM-powered decision engine utilizing ChatGPT for trade selection.
• Performance tracking via CSVs detailing daily PnL, equity, and trade history.
• Visualization tools using Matplotlib to compare performance against an index (e.g., Russell 2K).

Maintenance & Community

The project is part of a personal blog experiment with weekly updates posted on the author's blog and SubStack. Contact is available via email for feedback or advice.

Licensing & Compatibility

The repository appears to be under an unspecified license. The README does not explicitly state a license, which may require clarification for commercial use or integration into closed-source projects.

Limitations & Caveats

The project is described as an experiment with a small initial capital ($100) and a limited duration (6 months). Performance is tracked weekly, and the effectiveness of ChatGPT in generating alpha is the primary question being investigated, implying uncertain outcomes.