Crypto Market Data — Realtime Feeds and a Tick-by-Tick Historical Archive Across Every Venue

The jump in files and instruments from 2025 on is the prediction-market capture: every Kalshi and Polymarket event contract is its own instrument and its own day file, so a single series can add thousands of files per day. The crypto venues grow steadily by listings; the counts per venue are on the exchange pages.

318,832 instruments with recorded files across 32 venues, February 2021 – September 2026 — by class:

All venues ranked, month by month and by instrument class →

Our recorders sit next to each exchange’s matching engine and subscribe to the raw public feed — not a REST poll, not a vendor’s candle stream. Every message is stored as it arrived, with the venue’s own timestamp and our receive timestamp in integer nanoseconds, in one message schema for every venue: order-book snapshots and incremental updates, trades with aggressor side, top-of-book where the venue publishes one, and on derivatives the mark/index, funding and liquidation streams. What a day file holds, message type by message type, is on the historical data page and in the market-data specification.

Because the archive is the recording, not an aggregation of it, any statistic on this site can be reproduced from the files: the daily turnover and trade counts on the exchange and asset pages, the 1-minute OHLCV and spread statistics in the analytics, the contract results on the prediction-market pages. Gaps are shown as gaps — a day without a file is unknown, never zero.

There is no separate "market data API" product to license: the archive and its analytics are reachable programmatically as they are. The instrument search API resolves symbols to instrument IDs; the 1-minute statistics route serves the rolling last 24 h as CSV without a key; purchased and sample days download from authenticated routes that also serve the CSV/Parquet exports; and the MCP server exposes search, coverage, samples and checkout as tools for coding agents.

For live data, self-serve realtime access streams Binance USDT-M and Binance Spot over WebSocket by the UTC day — captured at the venue in Tokyo and delivered at our London, Frankfurt or Ashburn endpoints over dedicated lines, from €39.20 per instrument, €79.20 per market (public-beta price, 20 % off the list price of €49 / €99 for every day bought during the beta); the connection guide is at /docs/realtime. Bulk historical pulls use the order download routes documented for agents on llms.txt — one archive per day or one resumable tar per date range, never one request per file. Third-party exchange APIs are not proxied here; for a venue’s own endpoints and limits, check the current docs of that venue.

1. Choose the path: an instrument-day file for one venue and symbol, a series-day bundle for a whole Kalshi/Polymarket series or Deribit option chain, or the realtime feed for live data.
2. Try it first: every venue has free full-day samples in the exact shop format, with the CSV/Parquet exports and the browser order-book replay.
3. Find the instrument on its exchange page, asset page or via the search API, open its calendar in the shop and pick the UTC days you need — €1 per instrument-day, no minimum.
4. Buy per day at checkout, or in bulk with credit packs (€100 → 140 credits · €250 → 375 credits · €1,000 → 1,750 credits); files download instantly and stay yours.
5. Read the file with the free Python reader from the Agent Skill, or take the ?format= CSV/Parquet export straight into pandas, DuckDB or Polars — see the format docs.

Purchased data is for your own use — backtesting, research, analytics and training your own models, in-house or in your own products; the trained models and derived data you build from it are yours. The raw files and tick streams themselves may not be redistributed, resold or published. The full terms are on the data license page.