Kalshi API — Markets · Order Book · Trades

Live data from Kalshi, the CFTC-regulated US prediction market. Market odds ranked by traded volume, full L2 order-book depth, and executed trades with a volume-weighted average price. All public data — no login, no API key, no account.

Three datasets, priced separately

Mode	What you get	Price
markets	Live odds, ranked by traded volume	$0.005 per market
orderbook	Every bid and ask level with size	$0.02 per snapshot
fills	Executed trades with aggressor side	$0.002 per trade

You only pay for rows you actually receive.

Markets

Ranked by volume, because "the top 50 Kalshi markets" should mean the ones people trade. Kalshi's own flat market listing is dominated by untraded multi-leg parlay legs — of the first 1,200 it returns, 64 had any volume at all. This actor reads the event-level API instead, where 627 of 803 markets sampled were actively traded.

⚠️ You get a ranked slice, not the venue. Measured 2026-08-05 by walking Kalshi's full event cursor: 83,280 open markets exist, and Max results tops out at 500. So even a maximal run returns the top 0.6% by traded volume.

That is the intended product — the other 99.4% is overwhelmingly untraded parlay legs, and this actor exists precisely to rank past them. But it is a slice, and you should know its size before you build a "complete market list" on it. To follow specific markets beyond the top 500, pass their tickers to orderbook or fills directly; those take any ticker, ranked or not.

Field	Description
ticker	Unique market id, e.g. KXELONMARS-99 — use this for the other two modes
question	Self-identifying question text (see note below)
subtitle	What a YES resolves to, when one question is split across outcomes
yesPrice, noPrice	Implied probability, 0..1
spread	Bid/ask gap
volume, openInterest	Contracts traded / contracts outstanding
volumeUnit	Always contracts here. Polymarket reports shares, which is the same kind of unit ($1 max payout each), so those two are comparable. Manifold's play money is not.
closeTime, status, resolvedOutcome	Lifecycle

On question: Kalshi splits multi-outcome questions into one market per outcome and gives them all the same title — 138 of 803 live markets shared their title with another market, so two rows would read identically at different prices. Worse, some titles arrive with the subject missing ("Will become President of the United States?"). This actor rebuilds the question from the event so every row identifies itself. ticker remains the unique key.

Order book

Full L2 depth, normalized to 0..1 probability. Kalshi quotes both sides as bids, so a resting no-bid at 0.17 is published here as a yes-ask at 0.83 — the book reads like any other exchange's.

bestBid, bestAsk, midPrice, spread, plus bids[] and asks[] with price and size.

A book with no bids and no asks is not returned and not charged. An empty book is indistinguishable from a broken parser, so this actor treats it as a failure worth telling you about rather than a row worth billing.

Trades

Field	Description
yesPrice, noPrice	The probability this trade printed at
count	Contracts (each settles at $1)
value	What the YES buyer paid
takerSide	Which side was the aggressor — the direction that moved the price
isBlockTrade	Negotiated off-book size
ts	Venue timestamp

Each trade carries a summary with count, contracts, volume, aggressor split, and a single volume-weighted average price. That single VWAP is honest here — verified across 600 live trades, a Kalshi market's no price is always exactly 1 - yes, so there is one price series. Block trades are excluded from the VWAP because they print away from the book.

⚠️ The VWAP covers exactly the trades you fetched, and trades come newest-first. It is not the market's lifetime average unless you asked for enough rows to reach the market's first trade. Measured on KXALIENS-27: the newest 100 trades average 0.0725, all 26,910 average 0.1207 — the same field, 40% apart. The summary therefore carries oldestTradeAt, newestTradeAt and truncated, so you can always see which window you are looking at. If truncated is true, raise Max results to go further back.

Limits, stated up front

• No wallet or transaction hash. Kalshi is a regulated exchange, not a blockchain. Wallet-level whale tracking is not possible here — that is a Polymarket-only capability.
• liquidity is always null. Kalshi's API returned 0.0000 for it on all 803 markets sampled, including one with 167k volume. Reporting 0 would claim a market has no liquidity when it has a deep book, so this actor reports nothing rather than something false.
• Snapshots, not a stream. Each run captures a moment.
• What a run can cost you. Max results bounds the WHOLE run, and at $0.002 per trade the maximum of 50,000 rows costs $100. The default is 100 rows ($0.20), so you have to type a big number to spend a big number — but the ceiling is high on purpose, because reaching a market's first trade needs it. Set the number you actually want before you run.

Use cases

Quant research and backtesting, odds aggregation, cross-venue arbitrage against Polymarket, market-making analytics, AI agents that need live probability data.

---

For the developer (not shown on the Store)

Thin wrapper over the private scraper-gateway; contains no scraping logic. Set Actor environment variables (secret): SCRGW_URL, SCRGW_SECRET. Pay-per-event events: market-returned ($0.005), orderbook-snapshot ($0.02), fill-returned ($0.002). Deploy: apify push --force.