What it is
OddsPapi is a sports betting odds API that treats prediction markets as first-class bookmakers. Polymarket and Kalshi sit in the same response as Pinnacle, bet365, DraftKings and FanDuel, as native bookmaker slugs, with every price already normalised to decimal odds and every market carrying the same identifiers across books. Delivery is a REST API plus a low-latency WebSocket feed for line movement.
Its relevance in a prediction-market directory is not more market data — it is a reference point. That argument is the vendor's own, and it is a good one.
Why a prediction-market reader cares
A Polymarket or Kalshi quote on its own is hard to judge. Prediction markets run thinner liquidity and lower limits than major sportsbooks, so a wide price can mean an edge or simply mean nobody is quoting size. Put the same event next to a sharp book and you can start separating the two: compare the price, use the lay side and order-book depth that prediction markets expose under exchangeMeta, and check the stake limits returned with each outcome. This is the benchmark layer that the rest of this directory's venue pages cannot give you on their own.
What the coverage actually looks like
The vendor's marketing says 350+ bookmakers across 60+ sports, and its own May 2026 comparison post measured 370 bookmakers live from the /v4/bookmakers endpoint, visible to any key including the free tier. The catalogue breaks down into three groups worth knowing: sharp books for reference pricing (Pinnacle among them), exchanges including Betfair Exchange, Matchbook and an Asian exchange network (4casters, Fair999, Orbit, Piwi247 and Sharp Exchange), and crypto books such as 1xBet, Stake, BC.Game, Rollbit and Cloudbet. Prediction markets are a small, separate category on top.
Unusually, the site documents its own latency in detail — maximum live delay and pre-game delay per bookmaker, measured from the moment a price changes to the moment it reaches the WebSocket feed, with the polling-versus-push distinction spelled out for each. On the review date the coverage table on that page rendered no rows at all, so the metric, the method and the vendor's own observation that slower web-based books hover around 4–10 seconds are readable, but a current per-book figure is not.
The Kalshi caveat
This is the most important limitation for prediction-market users, and it is stated in the vendor's own developer guide rather than hidden: Kalshi appears as a bookmaker with live odds, but only for sports markets where coverage overlaps — the guide names NFL and NBA. Kalshi-only categories such as elections, economic indicators and weather are not covered and still need Kalshi's own API. Polymarket coverage is broader and includes the lay side, but the practical consequence is that OddsPapi is a bridge between prediction markets and sportsbooks, not a replacement for either venue's API.
Read-only by design
Every endpoint in the documentation is a read. OddsPapi aggregates, normalises and streams data; it does not place, modify or cancel orders, and it holds no funds. Execution stays at the venue, which means position sizing, fees, spreads and settlement risk remain yours to manage — the API only tells you what is being quoted.
Where it sits in this directory
Three layers are useful to keep apart. Venue APIs and pages tell you what a specific market is pricing. Dune gives you the authoritative historical record with SQL. OddsPapi sits between them as the cross-venue price feed: it answers what the same event is trading at elsewhere, including at books whose pricing is generally more efficient than a thin prediction market's. Alphascope-style research workspaces interpret; OddsPapi supplies.
Counters you should not quote
The site's bookmaker count still depends on which page you read, which is why none of the figures should be repeated as a fact: the documentation intro says 300+ bookmakers across 60+ sports, the pricing and bookmaker pages say 350+, the Polymarket bookmaker page repeats 350+, a US-locale hero recorded earlier in September read 125+ across 21+ sports, and the vendor's own May 2026 blog post measured 370 from the /v4/bookmakers endpoint. The difference is plausibly regional filtering of the catalogue, but the pages do not label the figures that way. The vendor's own recommendation is the right one: take a free key and call the bookmakers endpoint to check the books you actually need before paying for anything.





































