trading-2028/polymarket-ai-trading

This is a research and paper trading toolkit for Polymarket, a prediction market platform where people bet on real-world outcomes. The system connects to live Polymarket data and runs automated strategies that simulate trading without moving any real money by default. The core strategy is called mean reversion, which is based on the idea that prices that drift far from their historical average tend to drift back. The system runs three versions of this idea simultaneously with different risk levels: Conservative (smaller bets, stricter entry rules), Moderate (middle ground), and Aggressive (larger simulated bets, looser entry rules). An optional AI layer can help score market quality when you provide an OpenAI key. All trade history is stored in a local SQLite database, and a web dashboard lets you watch signals, model performance, and AI scoring panels in real time. You would use this if you want to study prediction markets, test trading ideas in simulation, or learn how automated market systems work without risking money. It is built for people comfortable with Node.js and Docker who want a structured research environment. The stack is Node.js with Express for the backend API, SQLite for storage, Docker for running all services together, and a static web dashboard that can be deployed to a hosting platform. No real wallet or funds are required to run the default paper trading mode.