Prediction Data
For Quants
High-fidelity orderbook snapshots from prediction markets. Don't trade on lag. Get the edge with millisecond-precision historical data for your backtests.
{ "id": "market_8x92", "probability": 0.65, "volume": 1250000 }
Everything you need to know
Common questions about our data infrastructure, simulation tools, and billing.
Prediction markets are volatile. Existing public APIs often miss high-frequency price movements or lack historical depth. PolyTape provides millisecond-precision snapshots, allowing you to backtest strategies on the actual liquidity that was available, not just the closing price.
We maintain direct websocket connections to prediction market orderbooks. We capture every update (bids, asks, and cancellations) in real-time and archive them into our high-performance ClickHouse database. This ensures you see the exact state of the book as it appeared during the event.
Our primary coverage is Polymarket (Polygon). Please note: We currently **exclude high-frequency 'Crypto Price' markets** (e.g., BTC/ETH 15m 1h 4h 1day up or down market). We are building dedicated, low-latency infrastructure to properly support these volatile markets, which will be released separately.
Our high-fidelity archival data is available starting from January 22, 2026. We do not backfill data prior to this date to ensure every snapshot is a verified, live-captured record from our websocket indexers.
Using mid-price is often inaccurate for larger trades. Our `/simulate/buy` and `/simulate/sell` endpoints solve this. You provide a timestamp and a trade size (e.g., $10,000), and our engine 'walks' the historical orderbook to calculate your effective entry price and slippage percentage based on the liquidity depth at that specific moment.
We use a simple Pay-As-You-Go model via USDC. 1 USDC = 1,000 API Credits. Credits never expire and are only deducted when you make a request. There are no monthly lock-ins, making it perfect for intermittent research or continuous trading.