An AI-powered event-driven pipeline automates the detection of breaking news, classifies its potential market impact using Claude, and executes trades on niche Polymarket markets. It targets users seeking an automated edge in prediction markets by leveraging real-time data and advanced AI classification, aiming for faster and more accurate trading signals than traditional bots.

How It Works

The V2 pipeline operates in an event-driven manner, ingesting news from real-time streams (Twitter, Telegram) with sub-5-second latency. It matches headlines to niche Polymarket markets (under $500K volume) and employs Claude to classify news impact as bullish, bearish, or neutral, alongside a materiality score. This classification-driven approach, coupled with edge detection and quarter-Kelly position sizing, enables automated trade execution when a profitable opportunity is identified in less competitive market segments.

Quick Start & Requirements

• Installation: A one-command setup is available via bash setup.sh. Manual setup involves cloning the repository (git clone https://github.com/brodyautomates/polymarket-pipeline.git), creating a Python virtual environment, installing dependencies (pip install -r requirements.txt), and configuring API keys in a .env file.
• Prerequisites: Requires Python 3, git, and API keys for Anthropic (required), Twitter (optional for real-time), Telegram (optional for real-time), and Polymarket (optional for live trading).
• Setup Time: Estimated at 2 minutes.
• Links: GitHub repository: https://github.com/brodyautomates/polymarket-pipeline.git.

Highlighted Details

• Real-time news detection via Twitter and Telegram streams, with RSS fallback.
• Claude AI for classifying news impact (bullish/bearish/neutral) and materiality.
• Focus on niche Polymarket markets with trading volumes under $500K to avoid high-frequency bot competition.
• Edge detection algorithm combined with quarter-Kelly position sizing for trade signal generation and risk management.
• Calibration tracking to monitor and report classification accuracy over time.
• Event-driven architecture utilizing asyncio for low-latency processing.

Maintenance & Community

The project is maintained by @brodyautomates. No other community links (Discord, Slack, etc.) are provided in the README.

Licensing & Compatibility

The README does not explicitly state a software license. This absence may imply a restrictive, non-standard, or unstated license, requiring clarification for commercial use or integration into closed-source projects.

Limitations & Caveats

The system operates in dry-run mode by default; live trading requires explicit configuration. Safety measures include a $25 maximum bet per trade and a $100 daily loss limit. The project is explicitly stated to be for entertainment and educational purposes only, not financial advice, and carries significant risk. Users should never trade with unaffordable funds, and past performance is not indicative of future results.