What are the rate limits on the Kalshi data API?

They scale by plan: the free Explorer tier is 1 request/second; Quant is 25 req/s; Desk Lite is 50 req/s and 3,000 req/min; Desk is 100 req/s and 6,000 req/min with dedicated capacity. The live WebSocket allows 1 to 100 subscriptions by plan. Per-plan X-RateLimit-* headers are returned on every response, so you can read your remaining quota in real time.

Can I backtest Kalshi strategies with this data?

Yes — that's what it's for. You get the full historical order book to replay a market and size fills against the liquidity that was actually there, then run the backtest in your own stack. Because history and the live stream share one JSON format, the code you backtest with also trades live.

Can I cancel anytime?

Yes. Paid plans are month to month and the free Explorer tier needs no card to start. Cancel anytime from the dashboard.

Kalshi data API

Kalshi Data API

One clean API for Kalshi depth: plain JSON over REST with a single Bearer key for history, and a live WebSocket stream for the present — both emitting the identical snapshot objects, so the code you backtest with reads the live feed unchanged.

The DepthFeed Kalshi data API serves order-book depth and price data as plain JSON over HTTPS. A metered REST API returns market discovery, metadata, and full-depth historical/latest snapshots with one Bearer key; a WebSocket stream pushes the same objects live. No SDK required — it drops into any stack, including function-calling and MCP-style agent tools.

Kalshi data API at a glance

Protocol: JSON over HTTPS — REST + WebSocket
Auth: Bearer / X-API-Key (or ?api_key= on WS)
Discovery: Markets → snapshots → live /v3/stream
Agent-ready: No SDK; MCP / function-calling friendly
Assets: 7 — BTC · ETH · SOL · XRP · DOGE · BNB · HYPE
Timestamps: Epoch-ms exchange + receive, per snapshot
Underlying price: Binance spot/futures, joined per snapshot
History: 7/30/90-day windows + full archive (Desk)
Delivery: REST API + live WebSocket, identical JSON
Resolution: Every change, or ?interval= 30s–1d downsample

What the API gives you

REST for history, WebSocket for live — one JSON shape

Discover markets, then pull full-depth snapshots for any market over the REST API. The live stream (wss://api.depthfeed.com/v3/stream) pushes the identical snapshot objects as books change, so the loader you write to replay history reads the live feed unchanged — backtest against the archive, then point the same code at the stream and trade.

Built for automation and agents

The API is plain JSON with one Bearer key (or X-API-Key), so it needs no SDK and drops straight into any function-calling or MCP-style tool definition. The WebSocket protocol is three JSON ops (subscribe / unsubscribe / ping), with ?api_key= auth where headers can't be set. Errors use a stable code enum you can branch on (AUTH_INVALID, COIN_NOT_IN_PLAN, RATE_LIMIT_BURST, …).

Metered, paginated, and tunable

Keyset (cursor) pagination and per-plan rate limits keep large pulls predictable, with X-RateLimit-* headers on every response. Choose your resolution: every recorded book change by default, or add ?interval= (30s–1d) to downsample server-side to one book per bucket. Full reference with copy-pasteable examples lives in the docs.

Start pulling kalshi data api

Free Explorer tier, no card. Full bid/ask depth and the underlying price on every snapshot, over a REST API and a live WebSocket stream.

Start free — no card Read the API docs

Questions, answered.

No. It's plain JSON over HTTPS with a single Bearer key, so it works from any language with an HTTP client and drops into function-calling or MCP-style agent tools without a wrapper. The live WebSocket uses three simple JSON ops (subscribe / unsubscribe / ping), and ?api_key= auth covers clients that can't set headers.