Download data
How to Download Polymarket Data
Downloading Polymarket data is a one-request move when you know the market slug: CSV for a first look, JSON for a pipeline, Parquet for analytics. This page walks the whole recipe, from finding the market to validating the rows you got.
Figures measured as of 2026-09-24 on the published PolyOrderbooks archive.
The recipe
Five steps to a downloaded market
Formats
Which format actually is best
CSV is the universal interchange and the honest first step for a single market; JSON keeps the nested ladder structure for code; Parquet is the right call beyond one market because it is columnar, compressed, and native to DuckDB and Polars — the stack this site's data libraries page ranks.
For a 5-minute market (~1,200 frames at 250ms), CSV is fine; for a month of a series, Parquet saves gigabytes and minutes.
Worked example
Download one market end to end
Find a resolved BTC 5-minute market slug, call /v1/markets/{slug}/books at resolution 250ms, and every row lands with timestamp, bids, asks, and depth. A 30-second research download now has the full odds of the market's life on one grid.
The same request with resolution 1s previews the shape for free while you stay inside Starter limits — which is the honest way to validate before pulling the full window.
Honest
The honest limits
The free tier's 3-day history is for tasting, not for shipping: production downloads need a window plan, and large Parquet lakebuilds belong in S3 exports. The recipe stays identical at every tier — only the window grows.
FAQ
Are there free Polymarket data downloads?
Yes — the free Starter plan exports any single market as CSV with 3 days of history; paid windows cover 30–120 days.
CSV, JSON, or Parquet?
CSV for inspection, JSON for nested ladder workflows, Parquet for analytical scale on the same data.
Can I download resolved markets?
Yes — resolved markets stay queryable with the winning outcome on every row; nothing is deleted after resolution.