Polymarket & Kalshi Price Data
Prediction-market price data starts with the contract's own 0–1 price and spread, derived from the recorded order book. Historical responses can also include the nearest preceding underlying crypto observation when one is available.
Prediction-market price data has two layers: the contract price and spread derived from a recorded book, plus a nullable ASOF underlying reference. Venue-specific millisecond observation times make the alignment inspectable while preserving each source's timing semantics.
Prediction-market price data, measured
Computed directly from our live capture across all three venues.
Polymarket + Kalshi + Limitless
archive starts January 2026
since January 2026
BTC · ETH · SOL · XRP · DOGE · BNB · HYPE
Measured directly from DepthFeed's live ClickHouse archive, June 21, 2026.
Anatomy of a real snapshot
This is one real Polymarket order-book snapshot, captured 10 milliseconds after the exchange stamped it; Kalshi's yes/no book follows the same shape. Every field is what the API returns.
Yes/Up token book; the No/Down side is the binary complement (down = 1 − up). — captured 10 ms after the exchange timestamp.
- price_up / price_down
- The market's own price — quoted 0–1 (implied probability). Down is the complement of Up.
- btc_price
- The underlying crypto reference price, ASOF-joined to this exact snapshot.
- orderbook_up.bids / .asks
- The full resting ladder — [price, size] on both sides — what a real order fills against.
- exch_ts_ms / recv_ts_ms
- Venue-specific millisecond timestamps; fields differ by source, so use the documented timestamp for each endpoint.
Polymarket & Kalshi price data at a glance
- Market price
- Outcomes quoted 0–1 (implied probability)
- Underlying
- Binance spot/futures, joined per snapshot
- Venues
- Polymarket · Kalshi · Limitless
- Capture
- Event-driven WS (Poly) · full-depth poll (Kalshi)
- Assets
- 7 — BTC · ETH · SOL · XRP · DOGE · BNB · HYPE
- Timestamps
- Millisecond observation time; venue-specific source stamps
- Underlying price
- Nearest preceding Binance reference when available
- History
- 7/30/90-day windows + full archive (Desk)
- Delivery
- REST history + normalized live WebSocket frames
- Resolution
- Raw stored observations or ?interval= 1s–1d
Two price layers, one timeline, every venue
The market's own price, from the real book
Each contract trades between 0 and 1, its price the market's implied probability. DepthFeed reads it from the live book — best bid, best ask, and mid at each change, not a stale last print. Polymarket carries a bid/ask book per outcome; Kalshi a yes/no ladder up to 100 levels per side; both arrive in the identical shape.
The underlying crypto price, joined
Where a preceding Binance spot/futures observation exists for the asset (BTC, ETH, SOL, XRP, DOGE, BNB, and HYPE), historical responses include it as a nullable ASOF reference. This supports alignment without claiming simultaneous updates.
One schema, any interval
The same JSON snapshot object describes Polymarket, Kalshi, and Limitless, so one loader reads every venue. Return every recorded change by default, or add ?interval= (30s–1d) to downsample server-side to one book per bucket.
Worked example: the same question, priced on both venues
The point of one schema is direct comparison. Here is the same BTC question on Polymarket and Kalshi.
1Both venues quote it
Ask 'will BTC be up over the next short window?' and both venues quote it — on Polymarket as an Up/Down token pair, on Kalshi as a yes/no contract. Both are 0–1, both are implied probability.
2Both quote tight
On a representative day, BTC short-dated markets quote a one-cent (single-tick) spread on both venues — Polymarket's Up token and Kalshi's yes contract alike — so you can compare the two books' probabilities directly.
3Read them on one timeline
Normalized price and depth concepts reduce venue-specific plumbing. Underlying references are nullable, and clients should preserve the documented timestamp semantics for each venue.
Precision notes — what's exact, and how we label it
Good data tells you how good it is. Here is what each price is, per venue.
Real bid/ask on every venue
The market price is the live best bid, best ask, and mid from the recorded book at each change — never a stale last print.
Settlement-exact where it counts (Polymarket)
Polymarket open/close references are captured exactly from the venue's own event metadata; 1h/4h/24h underlying is exact Binance, while 5m/15m intra-window uses a clearly-labeled Binance proxy.
Full-depth Kalshi yes/no
The complete Kalshi ladder is captured up to 100 levels per side, so every probability comes with its resting size attached.
Captured live, because depth can't be backfilled
Order-book history cannot be reconstructed after the fact, so we record it continuously — 640 million snapshots and counting, the deepest reaching back to January 2026.
Inspect the data before integrating
Start with a bounded sample, verify the fields and timestamps, then choose the API or research workflow that matches the job.
Questions, answered.
When available, historical responses include the nearest preceding Binance spot/futures observation. The reference is nullable and timestamped, so clients can measure and handle alignment gaps.