Data
Polymarket historical data
Marketlens keeps the historical record for 604,843 Polymarket markets across 644 recurring series and $3.3B of traded volume: L2 order books, executed trades, OHLCV candles, and resolution outcomes, a complete archive since March 2026. This page is an overview of the whole dataset; the order book specifics live on order book data, endpoint reference in the docs, and provider comparisons on compare.
What historical data exists
Four linked record types per market. Order books: full L2 snapshots about every 60 seconds, with every individual price change between them at millisecond resolution, so the book at any instant reconstructs exactly. Trades: every execution with price, size, side, and timestamp. Candles: OHLCV aggregated from those trades. Resolutions: once a market settles, its winning outcome is stored, currently for 600,816 resolved markets (the angle covered in depth on resolution history).
The archive is complete from March 2026 onward, not a rolling window, and every tier, including free, can query all of it.
Coverage by category
Every recurring Polymarket market type is recorded. Browse each category with per series market counts, resolution splits, and date ranges:
The full index, grouped by category with live coverage numbers, is the data catalog.
How to access it
Three paths to the same archive: the REST API (JSON with cursor pagination), the Python SDK (pip install marketlens, with pandas DataFrame helpers), and bulk Parquet exports per market or per series.
from marketlens import MarketLens
client = MarketLens()
# resolved markets with their outcomes
for market in client.markets.list(
category="Crypto", status="resolved", take=5,
):
print(market.question, market.winning_outcome)
# replay any order book tick by tick
for market, book in client.orderbook.walk(
"btc-up-or-down-5m",
after="2026-07-01T00:00:00Z",
before="2026-07-01T01:00:00Z",
):
print(book.as_of, book.bids[0], book.asks[0])The same data feeds the SDK's backtester, which replays real books with queue priority, latency, and fees.
How providers keep history
marketlens
polymarketdata.co
depthfeed.com
History depth and capture method vary widely; the full field is on the comparison page, with per vendor breakdowns linked from it.
FAQ
Common questions
Where can I get Polymarket historical data?
Marketlens serves the historical record for 604,843 Polymarket markets across 644 recurring series: L2 order books, executed trades, OHLCV candles, and resolution outcomes, through a REST API, a Python SDK (pip install marketlens), and bulk Parquet exports. The free tier includes 5M events per day with full archive access.
How far back does Polymarket historical data go?
Order book capture started in March 2026 and the archive is complete from there, not a rolling window. Every tier, including free, can query all of it. Each market's metadata reports its earliest captured snapshot.
What kinds of Polymarket historical data exist?
Four linked record types per market: full L2 order book snapshots, the stream of price changes between them at millisecond resolution, executed trades, and OHLCV candles. Once a market settles its winning outcome is stored too, currently for 600,816 resolved markets.
Does Polymarket have an official historical data API?
No. Polymarket's CLOB API serves the current book and live streams, not history. To study past markets you need a service that has been recording them; Marketlens has recorded every recurring Polymarket market continuously since March 2026.
Is Polymarket historical data free to download?
The Marketlens free tier includes 5M events per day with full API and full archive access, no card required. Paid plans start at $29/mo and raise rate limits and daily event budgets.
Try it
Query the full archive in two minutes
The free tier includes 5M events per day with full API and full archive access, no card required.
$ pip install marketlens