Glypse
Blog

Kalshi Market Data: What It Contains and How to Use It

Printed candlestick chart showing prediction market data

Kalshi market data is a set of fields attached to every contract: bid and ask prices, volume, open interest, price history and settlement rules. Kalshi publishes all of it through its API. Reading it well means pairing those fields with context: related markets on other venues, the news behind a move, and social sentiment. That pairing is what Glypse is built for. Its Research Pages put cited event briefings, price history and comparable contracts in one workspace, and the core tools are free without an account.

Key Takeaways

  • Each Kalshi contract exposes YES and NO bid and ask prices, volume, open interest, status and settlement rules, plus candlestick history.
  • Public market data needs no authentication, but REST responses can lag by seconds, so live work belongs on WebSockets.
  • A contract's price reads as an implied probability, and research on Kalshi's macro markets found those probabilities update quickly to news.
  • Volume, spread and open interest tell you how much weight a price deserves.
  • Cross-venue comparison, sentiment and AI-agent access add context that raw feeds do not include.

What Kalshi market data contains

Kalshi's Get Markets endpoint returns, for each contract, the YES and NO bid and ask prices in dollars, the size sitting at the best YES bid and ask, the last price and previous-day price, total and 24-hour volume in contracts, open interest, notional value at settlement, market type (binary or scalar), status and settlement rules.

A few terms in plain language:

  • Contract price: each contract settles at $1.00 or $0.00, so a price of $0.62 is read as roughly a 62% implied probability. Our guide to prices and probabilities walks through the conversion.
  • Volume: how many contracts traded, in total and over the last 24 hours.
  • Open interest: contracts currently outstanding, a measure of how much money is committed rather than how much changed hands.
  • Spread: the gap between the best bid and best ask. A narrow spread means you can enter and exit near the quoted price.

One detail surprises newcomers. Kalshi's orderbook API returns only bids, not asks, because a YES bid and a NO bid are reciprocal in a binary market. A YES ask is derived from the best NO bid. If you build your own tools, account for this. For the mechanics, see our guide to order books.

Getting the data through the API

Public market data for series, events, markets and orderbooks is available without authentication at the base URL given in Kalshi's market data quick start. Two practical points from that guide:

  1. REST responses can lag by several seconds. Kalshi recommends WebSockets for current prices and state.
  2. WebSocket connections require authentication.

Kalshi's documentation overview describes two API families, one for event-contract markets and one for perpetual futures, each reachable over REST, WebSocket and FIX. It also lists an MCP connector so AI assistants can explore markets, read portfolios and manage price alerts.

Filtering and paging

The Get Markets endpoint filters by status (unopened, open, closed, settled) and by timestamps for created, closed, settled and updated times. It also supports multivariate-event filtering. When you page through listings sorted by close time, the docs recommend deduplicating by ticker so a market is not counted twice.

Price history and its limits

The candlestick endpoint returns historical bid, ask and traded prices, volume and open interest. A single request can span at most 5,000 periods, so one-minute candles cover only about 3.5 days per call. Longer ranges must be split into several requests.

Older data lives elsewhere. Kalshi's historical data guide explains that live endpoints hold current and recent data, while older markets, trades, orders and candlesticks move to separate historical endpoints. A cutoff timestamp tells you where the split falls, and markets settled before it are reachable only through the historical endpoint. A backtest that ignores this will silently miss data.

Live updates

The Market Ticker channel sends one message per changed market with its current values and skips intermediate values. For every trade or each orderbook change, use the trade or orderbook delta channels instead.

SDKs

Kalshi publishes official Python and TypeScript SDKs, released roughly weekly to follow the API spec. The docs warn they may lag the API and recommend treating the OpenAPI and WebSocket specs as the source of truth.

Turning fields into probabilities

The academic case for reading prices as forecasts is solid. A January 2026 NBER working paper evaluates Kalshi's macro markets and finds well-calibrated, quickly updating forecasts for CPI, payrolls, unemployment, GDP and Fed funds rates. Its forecast errors on core CPI and unemployment were statistically similar to the Bloomberg consensus. The authors note that each contract pays $1 if its outcome occurs, so the contracts in one market together trace a full probability distribution.

The same paper gives an example of speed: the implied probability of a July 2025 rate cut rose to 25% after remarks from two Fed governors, then fell after a stronger-than-expected June jobs report.

The logic behind it comes from Chen and Pennock's survey of market design. A trader who believes the true probability is higher than the price can buy, and one who believes it is lower can sell, so the price settles near a consensus of what participants know.

Turning fields into probabilities

A worked example

Suppose a contract on a monthly CPI threshold trades with a YES bid of $0.58 and an ask of $0.62, and last week's close was $0.45. The midpoint, $0.60, implies about a 60% probability, up 15 points. Now ask what moved it. Was there a data release, a speech, a burst of volume? If volume jumped alongside the price, the move carries more weight than one made on a handful of contracts. Our guide to reading a market covers this routine in more detail.

What the numbers say about reliability

SignalWhat it tells youWatch for
Bid-ask spreadCost of entering and leavingWide spreads make the midpoint a rough estimate
24-hour volumeCurrent attentionOne-off spikes around news
Open interestCapital committedThin interest means a few traders set the price
Price historyHow fast beliefs changedGaps from the historical cutoff
Settlement rulesExactly what resolves YESWording differences between similar contracts

Read settlement rules before comparing any two contracts. Two markets with similar titles can resolve on different data sources or dates.

Comparing Kalshi data with other venues

Kalshi is large. Castle Labs research covering June 22 to 28, 2026 reports Kalshi with over 65% of prediction-market volume. In that window its BTC one-day binary market averaged $4.24 million in volume, against $2.76 million on Polymarket. Spread behavior also differed: Kalshi held a steadier band of 107 to 154 basis points, while Polymarket's reconstructed spread was tighter at 5.8 to 19.8 basis points.

Kalshi and Polymarket volume and spread comparison

The lesson for researchers is that the same event can have different depth, spreads and prices on different venues. When you compare, normalize for contract wording, fees and liquidity. Glypse's Cross-venue Search finds related contracts on Kalshi and Polymarket together, and the Arbitrage Scanner flags potential price gaps between them. For background on how venues differ, see prediction market platforms.

Analytics tools and alternative data

Raw Kalshi data answers what the price is. Analysis tools answer why it moved and what else is relevant. The useful categories are:

  • Screeners that filter thousands of contracts by category, price, volume or status. The Glypse Market Screener does this across Kalshi and Polymarket.
  • Event briefings that summarize the background of a contract with cited sources, so you can check the reasoning yourself.
  • Price-history analysis for spotting when and how a probability shifted.
  • Social sentiment as an alternative data stream. Glypse's Social Intelligence dashboard analyzes social media posts for sentiment, emotion, conversation volume, hashtags and cashtags, and trend changes. A rise in conversation volume ahead of a price move, or a divergence between sentiment and price, are patterns worth investigating.
  • Agent access through an MCP server, which lets an AI assistant query market data and research directly.

Where Glypse fits

Glypse combines these pieces in one workspace for Kalshi and Polymarket traders: cross-venue search, contract screening, AI-generated event briefings with cited sources, price history, arbitrage scanning and social sentiment. Core research tools are free without an account. Registration adds expanded results, saved queries, longer feeds, bookmarks and MCP access. Start from Research Pages with any contract that interests you, or read our guide to trading prediction markets first.

Using AI agents with market data

The Glypse MCP Server gives ChatGPT, Claude and other AI agents access to Glypse's market data, analytics and research. When an agent summarizes a market, check the figures it quotes against the contract's live price, volume and settlement rules, and confirm each claim against the cited source.

A simple research routine

  1. Find the contract and any equivalents on other venues.
  2. Read the settlement rules and close date.
  3. Check spread, volume and open interest to judge how much the price means.
  4. Review price history around known events.
  5. Look at sentiment and conversation volume for context.
  6. Save the query and revisit as new data arrives.

FAQ

Is Kalshi market data free?

Public data for series, events, markets and orderbooks is available without authentication through Kalshi's API. WebSocket streaming requires authentication. Glypse also offers its core research tools free without an account.

How far back does Kalshi price history go?

Live endpoints hold current and recent data, and older data sits in separate historical endpoints split by a cutoff timestamp. Each candlestick request is capped at 5,000 periods, so long ranges need several calls.

Why does the Kalshi orderbook show only bids?

In a binary market a YES bid and a NO bid are reciprocal, so the orderbook API returns bids on both sides. A YES ask is read from the best NO bid.

How do I compare Kalshi and Polymarket prices?

Match the contracts on wording, resolution source and date, then compare prices alongside spread and volume. Cross-venue search and an arbitrage scanner automate the matching.