Use cases & endpoints
Kalshi historical data — event contract price history
Event contracts turn a question into a price, and the price history is a record of how expectations moved as news arrived. That makes them genuinely useful for research — and easy to misread, because the number only means something alongside the exact resolution wording it settles against.
Two markets can look like the same question and resolve differently on a technicality. A contract about an economic release may hinge on which vintage counts; a contract about a decision may hinge on the announcement date rather than the effective date. Any dataset that stores only the price is discarding the part that determines what the price meant.
Hindcast keeps contract metadata and resolution criteria verbatim alongside timestamped price and volume observations, plus the settled outcome once a market resolves. Historical queries return what was quoted at the time, never a probability reconstructed with knowledge of the answer.
hindcast — session
previewhindcast> price history for an event contract before it settled
→ tool: prediction_market { venue:"kalshi", market_id:"...", history:true }
✓ timestamped prices and volumes with the resolution terms
source: venue market data and contract specification
What's covered
- Contract metadata with resolution criteria kept verbatim
- Timestamped price and volume observations
- Settled outcomes recorded when markets resolve
- Market lifecycle events — creation, close, settlement
- Links to related entities and topics where defensible
- No back-filled or smoothed probabilities
Join the waitlist — this page's dataset is already selected.
Questions
Can I compare a Kalshi market to a similar one elsewhere?
You can line them up, and the data keeps both sets of terms so you can see whether they are actually the same question. They frequently are not, and the difference in resolution wording usually explains the spread.
Are settled markets kept?
Yes, with their outcomes. Resolved markets are the only ones where forecast quality can be evaluated, so removing them would discard the most analytically useful part of the dataset.
Related
Join the waitlist. Vote the roadmap. First in, first served.