Real-Time Prediction Market Data.
Sub-second WebSocket and REST feed — Kalshi and Polymarket, normalized.
Live orderbook, trade prints, OHLC candles, ticker streams, and a heat-ranked featured list of what is moving right now — across both venues, through a single real-time prediction market data API. Built for agents, quants, dashboards, and research workflows.
First arrival — the wonder the Dutch felt landing in 1626, applied to discovering 9000+ live prediction markets.
Who uses the real-time data API?
Trading agents, dashboards, quants, backtest researchers, and risk engines — each subscribing to a different slice of the same normalized live feed.
Trading agent / quant bot
Subscribe to ticker + orderbook for active positions, run signal logic on tick callbacks
wss://data.simplefunctions.dev/v1/wsMarket dashboard / app
Featured top-N + per-market WebSocket for chart + book panel
GET /v1/markets/featured + WSBacktest researcher
Pull historical candles by timeframe; reconstruct tick replay from trade prints
GET /v1/candles + GET /v1/tradesRisk engine
Poll heartbeat + featured every minute; alert on heat regime shift
GET /v1/heartbeat + GET /v1/markets/featuredAI world-model loop
Featured stream as compact context for an LLM agent — what is hot right now
WS subscribe featuredREST endpoints
Edge-cached, normalized, with precomputed heat. All paths under thehttps://data.simplefunctions.dev/v1/base.
GET /v1/marketsAll tracked markets with precomputed heat score
max-age=3, swr=10GET /v1/markets/featured?n=20Top-N by heat — what is moving right now
max-age=3, swr=10GET /v1/markets/{ticker}Single market snapshot with full metadata
max-age=2, swr=5GET /v1/orderbook/{ticker}Live depth — bids, asks, timestamp
max-age=1GET /v1/candles/{ticker}?tf=1h&limit=500OHLC candles for any timeframe
max-age=15 (1d) / 5 (others)GET /v1/trades/{ticker}?limit=50Recent trade prints, newest first
max-age=1GET /v1/heartbeatSystem status — markets tracked, ws clients, uptime
max-age=10WebSocket topics
One connection,wss://data.simplefunctions.dev/v1/ws, fans out to any combination of topics. Frames are JSON.
ticker:{ticker}{type, ticker, last, bid, ask, volume24h, venue}On change
orderbook:{ticker}{type, ticker, bids, asks, ts}Throttled 1 Hz
trade:{ticker}{type, ticker, price, size, side, ts}On every print
candle:{ticker}:{tf}{type, ticker, timeframe, candle}On bar update
featured{type, markets, generated_at}60 s tick + rank-change
Heat scoring
A single 0–100 number per market that combines liquidity, contestation, and recent volatility. Featured rankings are heat-sorted; you can also re-sort client-side from the per-market heat field.
Volume
0.40ln(volume + 1) / ln(1e7 + 1) — log-scaled 24 h volume
Spread
0.351 − (spread / 0.05) — tighter spread = hotter
Close to 0.5
0.15Markets near coin-flip get a boost
Uncertainty
0.10Recent probability volatility
Featured response — example
Top-N markets by heat across both venues. Re-broadcast on thefeatured WebSocket topic when rank composition changes.
GET https://data.simplefunctions.dev/v1/markets/featured?n=20JSON · 200{
"generated_at": "2026-04-30T07:50:42Z",
"markets": [
{
"ticker": "KXFEDDECISION-26DEC-CUT100",
"venue": "Kalshi",
"title": "Fed rate cut > 100 bps by Dec 2026",
"yes_price": 0.32,
"bid": 0.31,
"ask": 0.33,
"volume_24h": 3284100,
"open_interest": 1520400,
"heat": 87.4,
"ts": "2026-04-30T07:50:41Z"
},
{ "...": "19 more, sorted by heat desc" }
]
}WebSocket quickstart
Subscribe to one or more topics in a single envelope. The server pushes JSON frames keyed by type.
JavaScript / TypeScriptbrowser + nodeconst ws = new WebSocket('wss://data.simplefunctions.dev/v1/ws')
ws.addEventListener('open', () => {
ws.send(JSON.stringify({
action: 'subscribe',
topics: [
'featured',
'ticker:KXFEDDECISION-26DEC-CUT100',
'orderbook:KXFEDDECISION-26DEC-CUT100',
],
}))
})
ws.addEventListener('message', (e) => {
const frame = JSON.parse(e.data)
if (frame.type === 'featured') {
console.log('top-N rank update', frame.markets.length)
}
if (frame.type === 'ticker') {
console.log(frame.ticker, frame.last)
}
})FAQ
What is the SimpleFunctions real-time prediction market data API?
A sub-second WebSocket and REST data feed for Kalshi and Polymarket prediction markets. You get normalized orderbook, trade prints, OHLC candles, ticker streams, and a heat-ranked featured list of what is moving right now — across both venues, through a single endpoint.
How does the WebSocket work?
Connect to wss://data.simplefunctions.dev/v1/ws and subscribe to topics like ticker:{ticker}, orderbook:{ticker}, trade:{ticker}, candle:{ticker}:{tf}, or featured. The server pushes JSON frames as data arrives. Orderbook frames are throttled to 1 Hz; trade frames are pushed on every print.
What is the heat score?
A single-number 0-100 ranking that combines log-scaled 24 h volume (40%), bid-ask spread tightness (35%), proximity to 0.5 probability (15%), and recent probability volatility (10%). Use it to find what is liquid, contested, and moving — not just what is popular.
How does this differ from the Kalshi or Polymarket APIs directly?
You get one endpoint that normalizes both venues, precomputed heat ranking, and a featured stream that updates on rank-change rather than every tick. Calling Kalshi and Polymarket directly works, but you write the normalization, heat scoring, and featured-rank logic yourself.
Can I subscribe to multiple tickers in one WebSocket connection?
Yes. The subscribe envelope accepts a topics: string[] field. One connection can fan out to dozens of markets simultaneously.
How fresh is the data?
REST endpoints have max-age caches between 1 and 15 seconds depending on the resource — orderbook and trades cap at 1 second. WebSocket frames are pushed on the venue tick rate (typically sub-second) with orderbook throttled to 1 Hz to keep frames lean.
What about historical data?
GET /v1/candles/{ticker}?tf=1h&limit=500 returns up to 500 OHLC candles per request across timeframes from 1 m to 1 d. GET /v1/trades/{ticker}?limit=50 returns the most recent prints. For deeper backtest pulls, contact us — we can stream older candles.
Is the data API paid?
A free tier is available with reasonable per-minute rate limits for testing and small agents. Higher-throughput tiers are coming — talk to us if you need more capacity right now.
How do I authenticate?
Free tier requires no API key for low-volume requests. Higher tiers issue an X-SF-API-Key header you attach to REST and pass as ?key= on WebSocket upgrade. The Vercel edge handles per-user rate limits and per-key quotas before requests reach the origin.
Does the data API include LLM-summarized state?
No — this API is the raw, machine-readable layer. For a 15-min summarized world snapshot for LLM agents, see /world. For event probabilities and meaning attached, see /event-probability-api. The realtime data API is the underlying tick layer that powers them.
What runtime powers this?
A Go origin on Fly.io ingests Kalshi and Polymarket continuously, computes heat in-process, and serves REST + WebSocket. The Vercel edge proxies via data.simplefunctions.dev, injecting an edge token, handling rate limits, and CORS.
Can I use this in a browser app?
Yes. CORS is enabled on data.simplefunctions.dev. Use fetch() for REST and the standard WebSocket API for streams. No special SDK required — but a thin TypeScript SDK is on the roadmap.
Related surfaces
Prediction Market API
Query, market, history, screen — Kalshi + Polymarket normalized for agents.
World State API
LLM-summarized 15-min snapshot — built on top of the real-time feed.
Event Probability API
Real-money event probabilities with gov/econ context attached.
Prediction Market Index
Live volatility, geo risk, breadth, activity gauges.
Prediction Market Hedging
Hedge real positions with binary event markets.
Kalshi CLI
Command-line trading and data tooling for Kalshi + Polymarket.
Institutions
Prediction market infrastructure for funds and corporates.