PuntersEdge API Review 2026

PuntersEdge is an Australian odds API. It reads 14 Australian-facing bookmakers and serves their prices over REST and JSON, with racing as the main event and sports odds alongside it.

The racing side is the reason to look at it. Thoroughbred, harness and greyhound racing get their own endpoint family: next-to-go, an upcoming race list, best price per runner, price history, steamers and drifters, settled results, and a permanent closing-line archive. Very few odds APIs cover Australian racing at all, and fewer still publish a closing-line archive for it.

Sports are a second product on the same key, and the sport list is longer than the pitch suggests: 20 of them, including AFL, NRL, NBA, MLB, NFL, NHL, tennis, cricket, MMA and the English Premier League. Depth follows the bookmakers rather than the league. On a full key we saw five AU books quoting head-to-head on NRL and AFL, two on the EPL, and one on MLB and NFL. Head-to-head is the market with real coverage. Spreads appear on three books and totals on two for NRL. There are no player props.

Authentication is a single X-API-Key header. Billing is credit-based, with the cost printed against each endpoint. There is a published OpenAPI 3.1 spec and a Python SDK on PyPI (pip install puntersedge).

The bookmaker list runs to Sportsbet, TAB, Neds, Ladbrokes, Unibet, PointsBet, Betr, Palmerbet, BetRight, TABtouch and others in the same bracket. The connector health endpoint shows 20 feeds ingested behind the 14 served books. Betfair Exchange and Pinnacle sit in that set as internal reference feeds and stay off customer keys, so do not plan around them. Bet365 is not in the set.

What stands out is that coverage is published as a measurement rather than a claim. A coverage report recomputes every 30 minutes and self-dates. At the time of writing it showed roughly 1,535 Australian races over the trailing 7 days with a median of 13 bookmakers quoting each race. That is a real number you can check, not a marketing figure, and it is a better signal about a provider than any feature list.

The freshness disclosure is equally unusual. Racing connectors report a median 7 to 13 seconds since last successful read for thirteen of the fourteen books, and 28 seconds for Betr. The provider publishes no single poll interval, on the grounds that the slowest served book runs four times slower than the fastest and one blanket number would be wrong for any named book. Every quote carries its own age_seconds, and /v1/data-quality/summary exposes per-connector freshness and audit status.

The closing-line archive reports itself the same way. Its coverage endpoint returned 927,122 rows across 25,647 races and 14 bookmakers, with 84% flagged as genuine closing lines. Results are collected for Australian and New Zealand racing from 15 August onward, and 94.2% of those 5,597 settleable races carry a finish position. Races elsewhere carry closing prices without a result, so the whole-table figure of 21% understates what you can actually settle. The endpoint returns both numbers, plus a results_scope note explaining the difference, which is a level of candour worth naming.

Worth knowing which side you are buying. Racing is where the speed is. Sports run on pre-match polling that tightens as a fixture approaches, so treat those prices as a pre-match reference rather than an in-play feed. Plan around that and the numbers hold up well.

Racing gets nine endpoints, and they are the well-designed part of the product:

• /v1/racing/next-to-go and /v1/racing/events for the card
• /v1/racing/best-odds for the top price per runner across books
• /v1/racing/price-history and /v1/racing/movers for line movement
• /v1/racing/results for settled races
• /v1/racing/closing-lines plus a companion /coverage endpoint that tells you what the archive actually holds

That last pair deserves credit. An archive endpoint that documents its own gaps is rarer than it should be, and it is what makes closing-line value work reliable rather than guesswork.

Sports add odds, historical snapshots and a movements feed, gated to the Standard tier and above. There are three arbitrage endpoints on top: a sports arb scanner, a best-prices comparison and a spreads/totals line arb.

Two details reward a closer look. Every racing quote carries a source_url pointing at that race on that bookmaker's site, which covered 100% of the quotes we pulled and saves you maintaining a link table. And every sports market carries a quality block with a score, an age in seconds, and a stale flag against a published threshold, so the feed tells you when to stop trusting it.

Delivery is REST plus outbound webhooks, with delivery logs and a test endpoint. There is no WebSocket and no server-sent events stream in the spec. For racing at a 7 to 13 second read cadence, polling is adequate. If your architecture assumes a push socket, plan on webhooks or a poll loop instead.

Two smaller touches are worth noting. An unrecognised bookmaker key returns a free 422 that names the valid keys, so a typo does not cost a credit. And the demo endpoints (/v1/demo/racing/next-to-go, /v1/demo/best-odds) run without a key, which means you can evaluate the response shape before you hand over an email address.

Five tiers, priced in Australian dollars, billed on monthly credits:

• Free: 1,500 credits/month, no credit card.
• Hobby, A$9/mo: 7,500 credits/month.
• Standard, A$29/mo: 75,000 credits/month. Unlocks historical snapshots and the movements feed.
• Business, A$99/mo: 300,000 credits/month.
• Unlimited, A$249/mo: 5,000,000 credits/month.

Credits are charged per call, and the rate varies: 1 credit for the sports list, 1 credit per market on an odds call, 3 credits for best-odds, 5 credits for a history or movements call. Work out your call pattern before you pick a tier, because a best-odds poll across several races burns credits three times faster than a raw odds pull.

At current rates A$29 is roughly USD 19, which puts the Standard tier at the cheap end of the market. The free tier at 1,500 credits is about 50 calls a day, enough to build against and not enough to run on. Paid plans go through Stripe checkout before the key is issued.

One licensing note: raw redistribution or resale of the data needs written permission. Applications, analytics and internal workflows are fine on paid tiers.

The product has a clear shape. Knowing the edges up front saves you a wasted integration:

• Bet365 and Pinnacle. Pinnacle is ingested as an internal reference feed and is not served to customer keys. Pair it with a Pinnacle-focused feed if you need a sharp reference price alongside the AU books.
• In-play sports odds. Sports are pre-match polling, not an in-play stream. The refresh tightens near the jump, but if you need true in-play prices this is not that feed. Racing is the fast side of the product.
• Player props and futures. Head-to-head carries the coverage. Spreads and totals exist on the larger AU books but thin out fast below them.
• WebSocket or SSE. Delivery is REST plus outbound webhooks, with delivery logs and a test endpoint.
• Stats and scores. You get the race card and settled results. Box scores, standings and season stats come from a separate source.
• MCP server. There is no packaged server for AI assistants, though the published OpenAPI 3.1 spec gives one plenty to work from.
• Overseas depth. The EPL, NFL, MLB and NHL are all served, but priced through Australian books, so you get one or two quotes per fixture rather than a market. Overseas race meetings work the same way.

The provider also documents that A-League and Big Bash are ingested and quoted by a served bookmaker, but return a 404 on customer keys. Publishing a gap like that instead of quietly rounding it into the coverage count says something good about how the rest of the numbers are reported.

If you are building anything that touches Australian racing, this is a narrow field and PuntersEdge is a serious entry in it. Next-to-go, per-runner best prices, movers and a permanent closing-line archive across 14 AU books is the right endpoint set for a racing tool, and A$29/mo for 75,000 credits is not much to pay for it.

It also suits anyone doing closing-line value work on AU racing. The archive plus the /coverage companion endpoint is the piece most providers leave out, and backtesting is guesswork without it.

It works well as a specialist leg of a stack. If you also need in-play sports odds, player props, or Bet365 and Pinnacle prices, run PuntersEdge for the AU racing and pair it with a broader feed for the rest. The free tier makes that easy to trial before you commit to either. For a broader feed to pair it with, browse the full API directory.

A specialist racing API that reports its own limits better than most providers report their strengths.

Australian racing across 14 bookmakers, a closing-line archive that documents its own holes, per-quote age fields, a self-recomputing public coverage report, and freshness published per connector instead of as one flattering average. Providers who publish numbers that make them look slower are usually the ones whose numbers you can trust.

Scope is the thing to weigh, and it is an easy call. Racing is the product and it is well built. Sports ride along on the same key as a pre-race reference, which is a useful extra rather than the reason to sign up. Buy it for the racing and the rest is upside.

The team also moves fast. We raised a handful of points during this review and they shipped changes, with a regression test behind one of them, inside a day.

The keyless demo endpoints return the same shape the paid tiers return, so you can check the response format and a few prices against your own books in about ten minutes. Then the free tier costs nothing to build against.

What stood out when we vetted PuntersEdge for the directory:

• Australian racing depth: thoroughbred, harness and greyhound odds across 14 AU bookmakers, from next-to-go to settled results
• Permanent closing-line archive with a coverage endpoint that reports what it holds, built for closing-line value work
• Published coverage and freshness numbers, per-quote age fields, an OpenAPI 3.1 spec and a Python SDK

The rare odds API built around Australian racing, and one that publishes its own coverage and freshness numbers instead of a marketing claim. Free tier with no card, keyless demo endpoints, and a team that shipped changes inside a day when we raised points during the review.

Last reviewed: 2026-09 · How providers get listed