Polymarket Orderbook Archive archive.pendulumflow.com

Contributed tier: devpyle

Contributed tier. Recorded and contributed by devpyle (github.com/devpyle); published here under CC-BY 4.0 at his request. 68 hours, 2026-08-15T10 to 2026-08-18T05 UTC, crypto Up-or-Down markets only (BTC, ETH, SOL, XRP, BNB, DOGE, HYPE; 5-minute, 15-minute and 1-hour markets). These hours predate this archive's own capture. The files are his format, not the v3 products: clob_l2 is a 5-level book snapshot at his receipt time (not the venue's full book with venue stamps), poly_trades are last-trade events, timestamps are int64 ms at his receipt (one NTP host, measured about 80 ms behind the venue, p95 230–800 ms), prices are float32 (compare with 1e-3 tolerance), rows are not de-duplicated across his recorder restarts. Every file matches his MANIFEST.txt by sha256 and size (verified 2026-09-16, 274/274). Nothing in this tier is merged with, or attested like, the witness products.

Licence and credit

Published under CC BY 4.0. Attribute it as devpyle (github.com/devpyle). This tier is contributed, not witnessed by this archive: we serve his bytes exactly as he uploaded them and have verified that they are the bytes he listed, nothing more.

Layout

contrib/devpyle/<date>/<date>T<hh>/<product>.parquet, 4 files per hour, 2026-08-15T10 to 2026-08-18T05, kept in the layout he recorded them in because his sums are over those bytes. At the top of the prefix: README.md (his, with the header above), LICENSE, MANIFEST.txt (his manifest, verbatim) and SHA256SUMS.txt (our re-measured sums of every served object). That checksum file is this tier's own ledger, separate from the v3 ledger. Browse the files.

Columns

ProductColumns
best_bid_askts, asset, token_id, best_bid, best_ask
clob_l2ts, asset, side, token_id, bid1–bid5, bid1_sz–bid5_sz, ask1–ask5, ask1_sz–ask5_sz, spread, imbalance
market_eventsts, event, market_id, condition_id, title, asset, token_ids, end_date
poly_tradests, asset, side, price, size, trade_id, is_buyer, token_id, market_id, maker_address, taker_address, fee_rate_bps

These are his column names, not the v3 schema. ts is int64 milliseconds at his receipt, not a venue timestamp; a touched price is not a fill, see A fill, from a price you touched.

Pendulum Flow Discord

Something wrong with an hour, or a file you need that is not here? Say so in the Discord, where the people who run the site and prediction market data geeks compare notes.

Join the Discord

Credit the collector: pendulumflow for V3, PMXT for V1 and V2, AG6 for their V2 archive. Ours and PMXT's are CC BY 4.0; AG6 states no licence. How to cite.

Serving these bytes is not endorsing them. We are not affiliated with Polymarket.

For AI readers: llms.txt, what this archive holds and the questions it cannot answer.

https://x.com/PendulumFlow

Join the Pendulum Flow Discord: other people who build on prediction-market data, comparing notes and sharing tooling and findings.