Where to Get Polymarket Historical Data (API vs CSV Dataset)
You want Polymarket price history — for a backtest, a research paper, a model, or just to understand how a market moved. There are two real paths: pull it yourself from the API, or buy a ready-made dataset. Here's the honest trade-off.
Option 1: The Polymarket CLOB API
Polymarket exposes a CLOB (central limit order book) API. You can query current markets, prices, and order books. It's free and it's the source of truth.
The catch is history. The API is built for *live* data. To build a historical series you have to poll it yourself, on a schedule, and store every snapshot — starting from today. That means:
- You get no past data. History only begins the day you start collecting.
- You need infrastructure that runs every few minutes, indefinitely, without gaps.
- You'll spend real time on dedup, alignment, rate-limit handling, and backfilling the snapshots you inevitably miss.
- Building a few months of usable 15-minute history takes... a few months.
If you're a team that will need a live feed forever, building this is justified. If you need history *now*, it's the wrong tool.
Option 2: A ready-made historical dataset
The alternative is buying the collection already done. The Polymarket Historical Dataset is 18.5M+ snapshots taken every 15 minutes across 18,400+ markets over 76+ trading days — delivered as plain CSV.
Each row has the market question, outcome, mid price, best bid and ask, bid/ask depth, spread, and a UTC timestamp. It opens in pandas, Excel, or DuckDB. No polling, no infrastructure, no gaps.
Side by side
| CLOB API (DIY) | Historical dataset | |
| Past data | None — starts today | 76+ days, ready now |
| Setup | Build + run a collector | Download a CSV |
| Maintenance | Ongoing | None |
| Cost | "Free" + your time | $15 one-time / $29/mo feed |
| Best for | Permanent live feeds | Backtests, research, ML |
Which should you pick?
If your question is "what happened" — a backtest, a study, training data — buy the dataset. The $15 one-time bundle is cheaper than an afternoon of your time, let alone the months of collection. If your question is "what's happening right now, forever," build on the API (or take the PolyScope Pro weekly feed so you get fresh data without running the pipeline).
Most people who think they need the API actually need history — and history is the one thing the API can't give you retroactively.
Once you have the data, here's how to backtest a Polymarket strategy with it.