This is an AI-powered forecasting system that analyzes prediction markets, identifies mispriced opportunities, and paper trades them autonomously.

Multi-Step Reasoning Pipeline

The current pipeline runs seven stages with six LLM calls (one Sonnet planner, one Sonnet synthesiser, three Opus forecasters in parallel, an Opus critic, and an Opus final integrator) plus a no-LLM research stage that fans out to ten conditional tools. The three perspective forecasters never see the market price; the price is revealed only at the critic stage. Every scan also runs a single-call zero-shot baseline on the same question, so we record six probabilities per scan and can ablate where the lift comes from at no extra inference cost.

Show previous V1 pipeline

Fullscreen

The Models

The V2 pipeline uses Anthropic models strategically. Claude Sonnet (fast, affordable) handles planning and evidence synthesis. Claude Opus (slower, more capable) handles the independent forecast, adversarial review, and final integration. Each forecast uses 2 Sonnet calls and 3 Opus calls (~$0.48 per forecast).

Paper Trading

All trades are simulated, no real money is at risk. The system tracks every trade as if it bought $1,000 worth of contracts at the market price when the edge was identified. P&L is calculated based on how the market moves after entry: if the agent bought YES at $0.40 and the market moves to $0.60, that's a $500 unrealized gain on $1,000 deployed. When the market resolves, winning YES contracts pay $1.00 each and losing ones pay $0.00.