This repository provides a paper-trading system for Polymarket, focusing on research and experimentation with prediction markets. It targets users interested in understanding prediction market dynamics, comfortable with Node.js and Docker, and seeking a simulated environment for strategy development without real financial risk. The system offers a dashboard for monitoring and analysis, enabling learning and simulated P&L tracking.

How It Works

The system ingests live data from Polymarket's CLOB API and processes it through three distinct mean-reversion trading models: Conservative, Moderate, and Aggressive. These models differ in their risk tolerance and signal confidence thresholds, employing Kelly-style sizing for simulated trades. An optional OpenAI integration allows for AI-driven market analysis and quality scoring. All historical data, signals, and model performance are stored in SQLite, accessible via a Node.js/Express API that powers a static web dashboard.

Quick Start & Requirements

• Primary install/run: Docker Desktop is recommended for the full stack. Navigate to the repository and run bash scripts/docker.sh start. For individual components, use npm run api or npm run dev:frontend.
• Prerequisites: Node.js 20+ (if not using Docker for all components), an OpenAI API key (for AI features), and a Bash environment (Git Bash, WSL) on Windows.
• Deployment: The backend can be hosted 24/7 on Render for approximately $7/month. The dashboard can be hosted on Vercel.
• Links: Live dashboard: polymarket-trading-dashboard.vercel.app. Backend example: polymarket-trading-system.onrender.com. OpenAI: platform.openai.com.

Highlighted Details

• Real-time data streaming from Polymarket's CLOB API.
• Three parallel trading "personalities" (Conservative, Moderate, Aggressive) based on mean-reversion strategies.
• Optional OpenAI integration for enhanced market analysis and scoring.
• A comprehensive web dashboard visualizing signals, model statistics, and AI insights.
• SQLite database for storing historical market data and trade logs, enabling comparison and analysis.

Maintenance & Community

This repository is a fork, originating from b1rdmania (upstream HKUDS-AI), with trading-2028 as the current maintainer. The project encourages forks and experimentation. Specific community channels like Discord or Slack are not detailed in the README.

Licensing & Compatibility

No explicit open-source license is stated. The project emphasizes "Paper trading & research," "Educational and research use," and encourages "Forks and experiments." While this suggests a permissive stance for non-commercial experimentation, the absence of a formal license like MIT or Apache means commercial use or closed-source linking requires explicit clarification.

Limitations & Caveats

This project is strictly for paper trading and research; it does not handle real money by default. It is not designed as a "plug-and-play" money-making solution. Live wallet trading, guaranteed profitability, or production-grade custody are explicitly out of scope and considered separate, high-risk endeavors.