Apify prediction-market scrapers
A store of third-party actors that normalize Kalshi and Polymarket into one row shape.
by Apify
Last updated
What it is
Not a product — a shelf. Apify hosts user-published programs called actors, and a crowd of authors has published near-identical ones that call Polymarket's and Kalshi's public APIs and emit one row per market in a shared shape: question, YES and NO price, best bid and ask, volume, liquidity, close time and a link back to the venue. A few also return order-book snapshots, resolved-market history or OHLC candles as separately priced events.
The scale of the duplication is the fact a reader most needs. In the Apify Store's own listing for prediction markets on 19 September 2026, 35 of the first 84 results named both Kalshi and Polymarket, published by 32 distinct accounts. Most of them had been created in the preceding weeks and carried between one and five users; the largest cross-venue installed base in that set belonged to an arbitrage scanner rather than to a feed. There is no leading actor here, no maintained reference implementation, and no reason to expect the one you pick to be the one that survives.
This card exists because the layer is real and readers will find it. It is not a recommendation of any actor in it.
Availability
Anyone with an Apify account, which needs an email address and no identity check. Apify's general terms, effective 9 July 2026, require account holders to be at least 18, state that the services are intended for business use and not for consumers, and require use consistent with sanctions and embargoes imposed by the Czech Republic, the European Union, the United States or the United Nations.
Nothing about the venues' own geoblocks applies here, because nothing is being traded: an actor reads public market data from a server, and where you may hold a position is a question for the venue's card, not this one.
Pricing
Two charges stack. The platform plan is Free at 0 USD with 5 USD a month of prepaid usage, Starter at 19 USD, Scale at 199 USD and Business at 999 USD, each carrying that much prepaid usage plus pay-as-you-go above it, cheaper compute units at the higher tiers, and a store discount that is applied to actor charges. Unused prepaid usage expires at the end of the billing cycle.
The actor then charges per event against that balance, and every author prices differently. Among the cross-venue actors checked on 19 September 2026, a market row ran anywhere between 0.00001 and 0.01 USD — a thousandfold spread for the same row out of the same public API — an order-book snapshot around 0.002 USD, and some actors add a flat charge per run (0.10 USD was one) or per actor start (0.02 USD was another). A full sweep of both venues is several thousand rows, so the arithmetic that matters is rows times rate times runs per day, and at the top of that range a quarter-hourly sweep is not a cheap way to hold a copy of two public APIs.
Markets & resolution
Whatever Kalshi and Polymarket list — in practice elections and politics, sports, crypto price
levels and a long tail of everything else. The actors resolve nothing, settle nothing and add
nothing to the resolution record beyond copying whatever rules text the venue exposes, so
resolution_source on this card is null.
The settlement field holds one value and the two venues do not agree: Kalshi settles in dollars,
Polymarket in USDC on-chain. Read cash here as a property of the sources, not of the scraper.
Integrations
Everything an actor produces lands in an Apify dataset, which is readable through the Apify API and downloadable as JSON, JSONL, CSV, XML, Excel, HTML table or RSS. Runs are startable from the web console, on a schedule, or through the API from Apify's own client libraries for Python and JavaScript. Actors are also exposed over the Model Context Protocol for agents, and the platform's MCP host answers an unauthenticated request with a 401 rather than a 404.
That plumbing is Apify's and it is the dependable part of this card. What varies is everything above it: input schema, field names, pagination behaviour and whether the run fails loudly or returns a short dataset when a venue's API changes shape.
Limitations
The support model is the limitation, and it is structural rather than a complaint about any one author. Apify's terms say that unless expressly stated otherwise, actors are not part of the services Apify provides, that the platform is offered as is with no representation as to suitability, availability, accuracy or completeness, and that a dispute with a third-party actor developer is yours to have with that developer. An actor whose author stops caring keeps its store page, keeps taking runs and keeps charging for them while returning less than it did.
There is no service level, no versioning promise and no changelog convention across this layer. The data is second-hand by construction — an actor reads the same public endpoints you could read, so anything the venue does not publish is not available here either, and anything the venue changes arrives here late.
For a one-off pull, a comparison table or a prototype, that is an acceptable trade for not writing two clients. For anything that has to be correct next month, the honest options are the venues' own APIs or a vendor whose name is on the data.
Alternatives
Polymarket's and Kalshi's public APIs are free, first-hand and documented by the people who run the matching engines; the work you avoid by using an actor is the normalisation between them. FinFeedAPI sells nine venues in one schema with a company behind it and usage-credit billing. Adjacent covers the same two venues as most of these actors and adds indices, reference rates and news.
Specs
- Interfaces
- API, Python, MCP server
- Export
- CSV, JSON, Xlsx
- Available in
- Global
- KYC required
- No
- Market subjects
- Politics, Sports, Crypto
- Resolved by
- Resolves nothing
- Maker fee
- None
- Platforms
- Web, Library
- AI features
- None
- Capabilities
- Order book, Cross venue
- Pricing verified
- Availability verified
Background
How this part of the sector works, rather than which product to pick.
- Where weather data comes from and what you may do with it — Observations, model output and archives are three products with three licences. The forecast is the cheap half; a clean observation history is not.
- Why the same contract costs two different prices on two venues — Two venues quoting one outcome are quoting two contracts. What the gap between the prices is made of - rules, fallbacks, fees, collateral, exits, eligibility.
- What a trade actually costs on a prediction market — Venues publish trading fees in four incompatible units, so 3.00% on one can be cheaper than 1.75% on another. How to convert them, and what else takes a cut.
Also worth comparing
- Adjacent — Prediction-market indices and reference rates, with Kalshi on the tier you can test.
- FinFeedAPI Prediction Markets API — Nine event venues normalized into one REST schema, billed by usage credit.
- Predictefy — Sixteen venues behind one verb family, priced in credits, with nothing readable for free.
- Kalshi API — REST, WebSocket and FIX access to a CFTC-regulated event exchange.
- Polymarket US API — A keyless public gateway for reading, and a signed key behind KYC for trading.
- DepthFeed — The recorded bid-ask ladder for crypto up-or-down markets, which no venue keeps itself.
FAQ
Is this one product?
No. It is a layer of separately published actors by unrelated authors, each with its own schema, pricing and maintenance. A search of the Apify Store for prediction markets on 19 September 2026 returned 35 actors naming both Kalshi and Polymarket, from 32 different accounts.
What does a run cost?
Each actor sets its own per-event price, and the spread is three orders of magnitude. Across the cross-venue actors checked on 19 September 2026 a market row cost between 0.00001 and 0.01 USD, with some also charging per run or per actor start, and the charges are drawn from the Apify plan's prepaid usage.
Does Apify stand behind these actors?
No. Apify's terms state that unless expressly stated otherwise, actors are not part of the services Apify provides, and that disputes with a third-party actor developer are between you and that developer.
Why use a scraper when both venues publish free public APIs?
The only thing being bought is the normalisation and the plumbing — one schema over two venues, pagination handled, output straight into a dataset your code can pull. If you are writing code anyway, the venues' own APIs cost nothing and do not go stale when an author loses interest.