ParlayAPI Review 2026

ParlayAPI is a credit-based sports odds aggregator with a catalog of 33 sportsbooks and data sources. The catalog spans US-regulated books (DraftKings, FanDuel, BetMGM, Caesars, Fanatics), sharps (Pinnacle), offshore (Bovada, bet365), DFS apps (PrizePicks, Underdog, Sleeper), prediction markets (Kalshi, Polymarket), and European books (Unibet, PMU). The list comes from a live /v1/bookmakers endpoint you can query yourself. The catalog describes sources, not uniform availability: check the sport, event, market and delivery route you need before you plan around it.

The API runs on a credit system rather than flat monthly pricing. One market in one region costs one credit. Free accounts get 1,000 credits a month, no card needed. Paid tiers run from $5 to $200 and multiply those credits into the millions. If you've used The-Odds-API, the request shapes look similar, which can ease a migration. Validate the actual parameters, event identities, odds formats and responses before you rely on that.

Data licensing covers personal analysis and internal tools, each person on their own account and key. Public display, shared feeds and resale are a conversation with ParlayAPI first: the Terms of Service and any written agreement govern. Read the Terms before you architect a product that re-serves the data.

The /v1/sports endpoint listed 373 sport keys when we checked. Keys describe what the catalog can address, not uniform availability: coverage varies by sport, event, market and delivery route. The big four US leagues (NFL, NBA, MLB, NHL) are there, plus NCAAF, NCAAB, WNBA, MMA, boxing, tennis (ATP, WTA, and dozens of tournament-level keys), golf, cricket, esports (League of Legends, CS2, Valorant), and 256 soccer competitions from MLS to the Premier League to regional leagues.

Market types include moneylines, spreads, totals, alternate lines, period markets (halves, quarters), and player props from 21 sources. Props are available from regulated books and from DFS platforms in the same endpoint, which is unusual. Futures (championship and season-long outrights) have their own endpoint.

Historical closing odds and match results reach back as far as 2005 for some datasets. Dates and completeness vary by dataset and by account access. The depth depends on your tier: the free plan looks back 48 hours, while the top-tier Scale plan gives you 10 years of access. A dedicated /v1/historical/sports/{sport_key}/closing-odds endpoint returns the final pre-game price per book, which is the number that matters for closing-line value (CLV) analysis.

The core API is REST with JSON responses. Auth goes in an X-API-Key header or a ?apiKey= query parameter. Every response includes credit-tracking headers (x-requests-used, x-requests-remaining) so you always know where you stand.

Streaming comes in two flavors:

• SSE (Server-Sent Events): available on Starter tier and above ($5/mo). The /v1/sports/{sport_key}/live/sse endpoint pushes play-by-play state changes. An Enterprise-only hot feed adds bookmaker and market filtering. Each SSE connection costs 5 credits, and a reconnect counts as a new connection, so keep it alive.
• WebSocket: available on Business tier and above ($40/mo). Two feeds exist: wss://parlay-api.com/v1/ws/odds/{sport_key} for full odds updates with arb flags, and /v1/ws/live/{sport_key} for a narrower game-lines-only stream. Latency runs 300 to 800ms on the Scale tier and up to 1 second on Business, per the docs.

The WebSocket sends four frame types: initial_state (a snapshot of the last 500 props on connect), odds_update (changed rows), arb_flagged (new arbitrage opportunity, scanned every 5 seconds), and heartbeat (every 30 seconds). Access is gated by tier, with no connection or per-frame credit charge on the odds feed.

Sandbox endpoints at /v1/sandbox/ serve synthetic data in the production schema with no key and no credits. They're the fastest way to wire up a client and verify the response shape before you spend anything.

Six tiers, all credit-based:

• Free, $0: 1,000 credits/month, 48-hour historical lookback, no card required.
• Starter, $5/mo: 20,000 credits, SSE streaming, 168-hour (1 week) lookback.
• Pro, $20/mo: 100,000 credits, 720-hour (30 day) lookback.
• Business, $40/mo: 1,000,000 credits, WebSocket + SSE streaming, 2,160-hour (90 day) lookback.
• Enterprise, $100/mo: 5,000,000 credits, 8,760-hour (1 year) lookback.
• Scale, $200/mo: 50,000,000 credits, 87,600-hour (10 year) lookback, 300-800ms WebSocket latency.

Rate limits are set per tier and published on the pricing page, so check there rather than treating any number here as a guarantee. Credits are consumed per call, and cost depends on the endpoint: odds calls cost markets times regions (so fetching three markets across two regions costs six credits), while props cost three credits flat and live endpoints cost one to five. A meta endpoint at /v1/meta/credit-costs gives you the exact formula for each route.

The free tier is genuinely useful for prototyping. It covers the core odds, events, scores, and props endpoints, and the sandbox routes need no key at all. When credits run out, paid endpoints return 403 while free endpoints stay accessible.

The scope is wide, but a few things sit outside it:

• Deeplinks. No links into the sportsbooks' apps or bet slips. The feed is read-only data.
• Bet settlement. No endpoint tracks whether a wager won or lost. Scores and results are available, so you can compute settlement yourself.
• Same-game parlays. No SGP pricing or correlation data.
• Team or player stats. Injury reports and news come through (sourced from ESPN), but box scores, advanced metrics, and standings are not part of the feed. Pair it with a stats provider from the directory if your app needs those.
• Public display, shared feeds and resale by default. The licence covers personal analysis and internal tools on your own account and key. For public display, a shared feed or resale, talk to ParlayAPI first; the Terms and any written agreement govern.

The credit-based model rewards developers who know which slices of data they need. A focused app that pulls NFL moneylines from three books costs a fraction of what a full-scan arbitrage engine burns, so pricing scales with actual usage rather than tier gates.

Three audiences get the most out of it:

• CLV and sharp-betting analysts who want historical closing lines from Pinnacle and other books going back years, without running their own archive.
• Prop bettors and DFS players who need PrizePicks, Underdog, and Sleeper lines alongside the regulated books in a single call.
• AI agent builders who want an MCP server with 22 tools, keyless sandbox endpoints for evaluation, and account actions (signup, login-link email, checkout link, saved preferences) that your MCP client should ask you to approve. The llms.txt and curated OpenAPI spec for tool-calling models are a nice touch.

If you're migrating from The-Odds-API, the similar request shapes give you a head start, but validate parameters, event identities, odds formats and responses against your own client. The credit model is different too, so map your current request volume to ParlayAPI credits before committing.

ParlayAPI packs a lot into a credit-based model: a 33-source catalog, hundreds of sport keys, closing-line archives that reach back years, and DFS and prediction-market sources that most aggregators skip. The free tier is real (no card, no expiry), the sandbox is genuinely keyless, and the MCP server shows the provider has thought about how AI agents consume data as well as human developers.

The tradeoffs are clear. WebSocket streaming is gated to Business tier ($40/mo) and above. The credit system means high-volume arb scanners need to budget carefully. And the licence covers your own analysis and internal tools; public display or resale needs a conversation with ParlayAPI first. None of those are hidden: the pricing page and the terms say it plainly.

The public surface area is unusually complete for a provider at this price point: sandbox routes, a live bookmaker count endpoint, transparent credit accounting, and a Terms of Service that spells out what you can and cannot build. Start with the sandbox, map your usage to credits, and the free tier will tell you whether the data fits.