{"id":3903,"date":"2026-09-15T10:00:00","date_gmt":"2026-09-15T10:00:00","guid":{"rendered":"https:\/\/oddspapi.io\/blog\/?p=3903"},"modified":"2026-09-08T13:39:58","modified_gmt":"2026-09-08T13:39:58","slug":"winamax-api-odds-access","status":"publish","type":"post","link":"https:\/\/oddspapi.io\/blog\/winamax-api-odds-access\/","title":{"rendered":"Winamax API: The .fr Suffix Adds 3.3 Points to Four Brands"},"content":{"rendered":"

Winamax is the biggest sportsbook in France and it has no public API. There is no developer portal, no documented endpoint, no key you can request. What Winamax does have in the OddsPapi catalogue is three slugs: winamax.fr<\/code>, winamax.es<\/code> and winamax.de<\/code>. All three carry live odds. None of them is a copy of the others.<\/p>\n

This post shows how to read all three in one call, and what a 69-fixture measurement says about which one you should trust. The short version: on the same fixture, at the same second, the French feed quotes a wider price than the German one. That holds on 49 of 49 fixtures where both ship a clean three-way.<\/p>\n

Then the pattern widens past Winamax. Six .fr<\/code> slugs sit on a Ligue 1 board, and each one prices its own domestic competition worse than its own international sibling does.<\/p>\n

Three Slugs, Two Feeds<\/h2>\n

The catalogue call is one line. On 8 September 2026 it returned 357 bookmakers, three of them Winamax.<\/p>\n

\n\n\n\n\n\n\n\n
Slug<\/th>\nName<\/th>\nliveOdds<\/th>\ncloneOf<\/th>\n<\/tr>\n<\/thead>\n
winamax.fr<\/code><\/td>\nWinamax FR<\/td>\ntrue<\/td>\nnull<\/td>\n<\/tr>\n
winamax.es<\/code><\/td>\nWinamax ES<\/td>\ntrue<\/td>\nnull<\/td>\n<\/tr>\n
winamax.de<\/code><\/td>\nWinamax DE<\/td>\ntrue<\/td>\nnull<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n

cloneOf<\/code> is null on all three, which tells you nothing. It is null on plenty of slugs that quote byte-identical prices. The only way to know whether two feeds are the same feed is to compare the numbers.<\/p>\n

So we did. Across 69 fixtures in seven competitions, winamax.es<\/code> and winamax.de<\/code> posted the same three-way tuple on 42 of the 45 fixtures where both were readable. winamax.fr<\/code> matched neither one, on any of the 49 fixtures readable against the German feed or the 42 against the Spanish. Three slugs, two pricing engines: a French one, and a shared international one.<\/p>\n

Old Way vs OddsPapi<\/h2>\n
\n\n\n\n\n\n\n\n\n\n\n
<\/th>\nScraping Winamax<\/th>\nOddsPapi<\/th>\n<\/tr>\n<\/thead>\n
Access<\/td>\nNo public API, geo-locked site, session tokens<\/td>\n?apiKey=<\/code> query parameter<\/td>\n<\/tr>\n
Regional feeds<\/td>\nThree separate sites, three scrapers<\/td>\nThree slugs in one call<\/td>\n<\/tr>\n
Comparison books<\/td>\nBuild one scraper per book<\/td>\n281 books on the same fixture payload<\/td>\n<\/tr>\n
Price history<\/td>\nYou store it yourself from day one<\/td>\n\/historical-odds<\/code> on the free tier<\/td>\n<\/tr>\n
Market coverage<\/td>\nWhatever the page renders<\/td>\n61 to 67 market IDs per Winamax feed<\/td>\n<\/tr>\n
Push updates<\/td>\nPoll and hope<\/td>\nWebSocket stream<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n

Step 1: Find a French Fixture<\/h2>\n

Authentication is a query parameter. There is no header, no OAuth dance, no bearer token.<\/p>\n

import os\nimport requests\n\nAPI_KEY = os.environ[\"ODDSPAPI_API_KEY\"]\nBASE = \"https:\/\/api.oddspapi.io\/v4\"\n\nfixtures = requests.get(f\"{BASE}\/fixtures\", params={\n    \"apiKey\": API_KEY,\n    \"sportId\": 10,          # soccer\n    \"from\": \"2026-09-08\",\n    \"to\": \"2026-09-16\",\n}).json()\n\nligue1 = [f for f in fixtures\n          if f[\"tournamentName\"] == \"Ligue 1\"\n          and f[\"categoryName\"] == \"France\"\n          and f[\"hasOdds\"]]\n\nfor f in ligue1[:3]:\n    print(f[\"fixtureId\"], f[\"startTime\"],\n          f[\"participant1Name\"], \"v\", f[\"participant2Name\"])\n<\/code><\/pre>\n
id1000003472036180 2026-09-11T18:45:00.000Z Stade Rennais FC v Olympique Marseille\nid1000003472036182 2026-09-12T15:15:00.000Z Strasbourg Alsace v AS Monaco\nid1000003472036166 2026-09-12T18:45:00.000Z AJ Auxerre v OGC Nice\n<\/code><\/pre>\n
Filter on categoryName<\/code>, because “Ligue 1” is also the name of the Algerian top flight. Two Algerian Ligue 1 fixtures sat in the same window while we ran this. The tournamentName<\/code> alone will hand you the wrong country.<\/p>\n
Set to<\/code> to the day after the last day you want. It is a midnight UTC instant, so a same-day range returns only the 00:00Z kick-offs.<\/p>\n
Step 2: Pull All Three Feeds in One Call<\/h2>\n
The bookmakers<\/code> filter takes a comma-separated list. Ask for all three Winamax slugs at once.<\/p>\n
FIXTURE = \"id1000003472036180\"   # Stade Rennais v Olympique Marseille\nWINAMAX = \"winamax.fr,winamax.es,winamax.de\"\n\nodds = requests.get(f\"{BASE}\/odds\", params={\n    \"apiKey\": API_KEY,\n    \"fixtureId\": FIXTURE,\n    \"bookmakers\": WINAMAX,\n}).json()[\"bookmakerOdds\"]\n\nprint(sorted(odds))\n# ['winamax.de', 'winamax.es', 'winamax.fr']\n<\/code><\/pre>\n
Step 3: Read the Three-Way and Measure the Margin<\/h2>\n
Soccer’s full-time result is market 101<\/code>, with outcomes 101<\/code> home, 102<\/code> draw and 103<\/code> away. Three flags decide whether a quote is real, and you have to check all three. The book-level suspended<\/code> flag and bookmakerIsActive<\/code> sit on the bookmaker object. The active<\/code> flag sits on each price. A parked board ships every price with active: true<\/code> underneath a suspended: true<\/code> book.<\/p>\n
FULL_TIME_RESULT = \"101\"\nLEGS = {\"101\": \"home\", \"102\": \"draw\", \"103\": \"away\"}\n\ndef read_1x2(book):\n    if book.get(\"suspended\") or book.get(\"bookmakerIsActive\") is False:\n        return None\n    market = book.get(\"markets\", {}).get(FULL_TIME_RESULT)\n    if not market or not market.get(\"marketActive\"):\n        return None\n    prices = {}\n    for outcome_id, name in LEGS.items():\n        outcome = market[\"outcomes\"].get(outcome_id)\n        if outcome is None:\n            return None\n        price = next((p for p in outcome[\"players\"].values() if p[\"active\"]), None)\n        if price is None:\n            return None\n        prices[name] = price[\"price\"]\n    return prices\n\ndef overround(prices):\n    return (sum(1 \/ p for p in prices.values()) - 1) * 100\n\nfor slug in sorted(odds):\n    book = odds[slug]\n    prices = read_1x2(book)\n    if prices is None:\n        print(f\"{slug:12} skipped: suspended={book['suspended']}\")\n    else:\n        print(f\"{slug:12} {prices}  margin {overround(prices):.2f}%\")\n<\/code><\/pre>\n
winamax.de   {'home': 2.1, 'draw': 4.0, 'away': 2.9}   margin 7.10%\nwinamax.es   skipped: suspended=True\nwinamax.fr   {'home': 2.1, 'draw': 3.95, 'away': 2.9}  margin 7.42%\n<\/code><\/pre>\n
The two live feeds differ by five hundredths on the draw and nothing else. That is enough to move the margin by a third of a point. Across the 69 fixtures, winamax.fr<\/code> shipped suspended: true<\/code> on 14, winamax.es<\/code> on 17 and winamax.de<\/code> on 7. Skip that check and you will average a parked board into your consensus.<\/p>\n
The players<\/code> dict is keyed by player ID on prop markets and by \"0\"<\/code> on game lines. Iterate the values rather than hardcoding players[\"0\"]<\/code>, and the same reader works on both.<\/p>\n
Step 4: Rank Winamax Against the Whole Board<\/h2>\n
Drop the bookmakers<\/code> filter and the same fixture returns everything the key can read.<\/p>\n
board_raw = requests.get(f\"{BASE}\/odds\", params={\n    \"apiKey\": API_KEY, \"fixtureId\": FIXTURE,\n}).json()[\"bookmakerOdds\"]\n\nboard = {}\nfor slug, book in board_raw.items():\n    if slug.startswith(\"pinnacle+\") or slug == \"demo\":\n        continue                       # internal test feeds\n    prices = read_1x2(book)\n    if prices:\n        board[slug] = prices\n\nranked = sorted(board, key=lambda s: overround(board[s]))\ntuples = {tuple(p.values()) for p in board.values()}\n\nprint(len(board_raw), \"books in the payload\")\nprint(len(board), \"quote a complete active three-way\")\nprint(len(tuples), \"independent price tuples\")\n\nfor slug in [\"pinnacle\", \"winamax.de\", \"winamax.fr\", \"bet365.fr\", \"bwin.fr\"]:\n    if slug in board:\n        print(f\"{slug:12} rank {ranked.index(slug) + 1:>3} of {len(ranked)}\"\n              f\"   {overround(board[slug]):.2f}%\")\n<\/code><\/pre>\n
281 books in the payload\n267 quote a complete active three-way\n87 independent price tuples\npinnacle     rank  30 of 267   3.48%\nwinamax.de   rank 187 of 267   7.10%\nwinamax.fr   rank 189 of 267   7.42%\nbet365.fr    rank 263 of 267   12.90%\nbwin.fr      rank 264 of 267   14.56%\n<\/code><\/pre>\n
267 books, 87 independent prices. Two thirds of that board is duplicate feeds quoting the same tuple, which is why you dedupe on the price tuple before you compute a consensus<\/a>. Counting slugs instead of prices will triple-weight whichever group happens to be largest.<\/p>\n
The .fr Premium Runs Across Six Brands<\/h2>\n
Look at the ranks again. bet365.fr<\/code> at 263 and bwin.fr<\/code> at 264, out of 267. The international bet365<\/code> slug on the same fixture prices at 10.11% and bwin<\/code> at 6.51%. Same brand, same match, same second.<\/p>\n
We ran the pairing across all 69 fixtures: every .fr<\/code> slug against its own brand’s international feed on the same fixture. Four brands ship both.<\/p>\n
\n\n\n\n\n\n\n\n\n\n
French feed<\/th>\nParent feed<\/th>\nPaired reads<\/th>\nMedian gap<\/th>\n.fr wider on<\/th>\n<\/tr>\n<\/thead>\n
bwin.fr<\/code><\/td>\nbwin<\/code><\/td>\n65<\/td>\n+8.81 pp<\/td>\n65 of 65<\/td>\n<\/tr>\n
bet365.fr<\/code><\/td>\nbet365<\/code><\/td>\n69<\/td>\n+3.02 pp<\/td>\n69 of 69<\/td>\n<\/tr>\n
unibet.fr<\/code><\/td>\nunibet<\/code><\/td>\n64<\/td>\n+3.01 pp<\/td>\n62 of 64<\/td>\n<\/tr>\n
winamax.fr<\/code><\/td>\nwinamax.de<\/code><\/td>\n49<\/td>\n+1.60 pp<\/td>\n49 of 49<\/td>\n<\/tr>\n
All four<\/strong><\/td>\n247<\/strong><\/td>\n+3.28 pp<\/strong><\/td>\n245 of 247<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n
Six .fr<\/code> slugs appeared on these boards: bet365.fr<\/code>, bwin.fr<\/code>, pokerstars.fr<\/code>, unibet.fr<\/code>, winamax.fr<\/code> and zebet.fr<\/code>. On the 18 French league fixtures their median overrounds ran 9.75% to 14.94%, against a board median of 7.20% and a Pinnacle median of 3.88%.<\/p>\n
The suffix predicts the price better than the brand does. If you are building a French-facing product and you reach for the French slug because it looks like the right one, you are picking the widest version of that book on purpose.<\/p>\n
PAIRS = [(\"winamax.fr\", \"winamax.de\"), (\"unibet.fr\", \"unibet\"),\n         (\"bwin.fr\", \"bwin\"), (\"bet365.fr\", \"bet365\")]\n\nfor local, parent in PAIRS:\n    if local in board and parent in board:\n        gap = overround(board[local]) - overround(board[parent])\n        print(f\"{local:12} {overround(board[local]):5.2f}%\"\n              f\"   {parent:12} {overround(board[parent]):5.2f}%\"\n              f\"   {gap:+.2f} pp\")\n<\/code><\/pre>\n
winamax.fr    7.42%   winamax.de    7.10%   +0.32 pp\nunibet.fr     6.68%   unibet        7.08%   -0.40 pp\nbwin.fr      14.56%   bwin          6.51%   +8.04 pp\nbet365.fr    12.90%   bet365       10.11%   +2.79 pp\n<\/code><\/pre>\n
One fixture is not the study. On this particular match unibet.fr<\/code> came in 0.40 points tighter than its parent, one of the two reversals in 247. Run the loop over a slate before you trust the direction, which is the whole reason we ran 69.<\/p>\n
Where the Gap Is Widest<\/h2>\n
The two Winamax feeds track each other closely in the top flights and separate in the second tiers.<\/p>\n
\n\n\n\n\n\n\n\n\n\n\n\n
Competition<\/th>\nFixtures<\/th>\nBoard median<\/th>\nwinamax.fr<\/code><\/th>\nwinamax.de<\/code><\/th>\nGap<\/th>\npinnacle<\/code><\/th>\n<\/tr>\n<\/thead>\n
Premier League<\/td>\n10<\/td>\n6.04%<\/td>\n7.70%<\/td>\n6.83%<\/td>\n0.79 pp<\/td>\n3.60%<\/td>\n<\/tr>\n
LaLiga<\/td>\n13<\/td>\n6.25%<\/td>\n7.90%<\/td>\n7.01%<\/td>\n1.06 pp<\/td>\n3.87%<\/td>\n<\/tr>\n
Serie A<\/td>\n10<\/td>\n6.22%<\/td>\n8.00%<\/td>\n6.93%<\/td>\n1.28 pp<\/td>\n3.70%<\/td>\n<\/tr>\n
Ligue 1<\/td>\n9<\/td>\n6.33%<\/td>\n8.12%<\/td>\n6.80%<\/td>\n1.29 pp<\/td>\n3.35%<\/td>\n<\/tr>\n
Bundesliga<\/td>\n9<\/td>\n6.41%<\/td>\n8.90%<\/td>\n7.23%<\/td>\n1.91 pp<\/td>\n3.68%<\/td>\n<\/tr>\n
Ligue 2<\/td>\n9<\/td>\n7.87%<\/td>\n14.87%<\/td>\n12.10%<\/td>\n2.17 pp<\/td>\n5.62%<\/td>\n<\/tr>\n
2. Bundesliga<\/td>\n9<\/td>\n7.60%<\/td>\n14.84%<\/td>\n11.90%<\/td>\n2.99 pp<\/td>\n4.56%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n
The ordering is consistent. Premier League 0.79 points apart, 2. Bundesliga 2.99. Both Winamax feeds roughly double their margin when they drop from a first division to a second, and the French feed doubles further. Pinnacle widens too, from 3.35% on Ligue 1 to 5.62% on Ligue 2, so some of this is the competition rather than the book. The difference between two feeds of one brand<\/em> is not a competition effect.<\/p>\n
If your model reads Ligue 2, this is the whole ballgame. A 14.87% overround against a 7.87% board median means you are paying twice the going rate for the same opinion. Ligue 1 alone carries 174 books a fixture<\/a>, and the second tier still returned 258 on the fixtures we sampled.<\/p>\n
Two Brands, One Price<\/h2>\n
The clone check turned up a pair that has nothing to do with Winamax and everything to do with your consensus maths.<\/p>\n
\n\n\n\n\n\n\n\n\n
Slug A<\/th>\nSlug B<\/th>\nIdentical three-way on<\/th>\n<\/tr>\n<\/thead>\n
unibet.fr<\/code><\/td>\nzebet.fr<\/code><\/td>\n66 of 66 fixtures<\/td>\n<\/tr>\n
winamax.de<\/code><\/td>\nwinamax.es<\/code><\/td>\n42 of 45 fixtures<\/td>\n<\/tr>\n
unibet<\/code><\/td>\nunibet.be<\/code><\/td>\n42 of 67 fixtures<\/td>\n<\/tr>\n
pmu<\/code><\/td>\nunibet<\/code><\/td>\n30 of 67 fixtures<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n
Unibet France and Zebet are two brands, two websites and two licences, and they posted the same numbers on every fixture we read. cloneOf<\/code> is null on both. Add them both to a fair-value average and you have weighted one opinion twice.<\/p>\n
The rule that survives every one of these audits: dedupe on the price tuple, never on the metadata. Post 3731 found the same thing inside Bwin’s seven regional feeds<\/a>, where the international slug turned out to share a price with BetMGM.<\/p>\n
When a Slug Returns Nothing<\/h2>\n
Winamax France stops at the top two divisions. On a French third-tier fixture the German feed priced it and the French one did not. The failure modes look identical from the outside until you probe one slug at a time.<\/p>\n
def probe(fixture_id, slug):\n    r = requests.get(f\"{BASE}\/odds\", params={\n        \"apiKey\": API_KEY, \"fixtureId\": fixture_id, \"bookmakers\": slug,\n    })\n    if r.status_code == 400:\n        return \"not a slug in the catalogue\"\n    if r.status_code == 403:\n        return \"real slug, gated on your plan\"\n    if slug in r.json().get(\"bookmakerOdds\", {}):\n        return \"priced\"\n    return \"real slug, no board on this fixture\"\n\nNATIONAL = \"id1000018372523912\"   # SM Caen v Paris 13 Atletico\n\nfor slug in [\"winamax.fr\", \"winamax.de\", \"winamax.co.uk\"]:\n    print(f\"{slug:14} {probe(NATIONAL, slug)}\")\n<\/code><\/pre>\n
winamax.fr     real slug, no board on this fixture\nwinamax.de     priced\nwinamax.co.uk  not a slug in the catalogue\n<\/code><\/pre>\n
Three distinct answers, three distinct status codes. The 400 body is worth reading on its own: it enumerates every valid slug, which makes it a free catalogue lookup when you have fat-fingered a name.<\/p>\n
Note the empty case. When a valid slug has no board on a fixture, the response omits bookmakerOdds<\/code> entirely rather than returning it empty. Use .get(\"bookmakerOdds\", {})<\/code> or your loop dies on a KeyError<\/code>.<\/p>\n
Free Winamax Price History<\/h2>\n
Historical odds are on the free tier, which is the part competitors put behind an enterprise contract. Point \/historical-odds<\/code> at the same fixture and you get every snapshot recorded since the board opened.<\/p>\n
history = requests.get(f\"{BASE}\/historical-odds\", params={\n    \"apiKey\": API_KEY,\n    \"fixtureId\": FIXTURE,\n    \"bookmakers\": \"winamax.fr,winamax.de,pinnacle\",   # max 3 per call\n}).json()[\"bookmakers\"]\n\nfor slug, book in history.items():\n    market = book[\"markets\"][FULL_TIME_RESULT]\n    snapshots = market[\"outcomes\"][\"101\"][\"players\"][\"0\"]\n    snapshots.sort(key=lambda s: s[\"createdAt\"])\n    changes = sum(1 for a, b in zip(snapshots, snapshots[1:])\n                  if a[\"price\"] != b[\"price\"]) + 1\n    print(f\"{slug:12} {len(snapshots):>3} snapshots,\"\n          f\" {changes:>3} distinct home prices,\"\n          f\" first {snapshots[0]['createdAt'][:10]}\")\n<\/code><\/pre>\n
pinnacle       51 snapshots,  38 distinct home prices, first 2026-08-30\nwinamax.de     69 snapshots,   7 distinct home prices, first 2026-08-31\nwinamax.fr     26 snapshots,   6 distinct home prices, first 2026-08-31\n<\/code><\/pre>\n
Collapse consecutive identical prices before you count anything. The feed records a snapshot on a schedule whether or not the price moved, so 69 raw snapshots became 7 repricings. Skip that step and you will measure your own polling interval.<\/p>\n
The endpoint caps at three bookmakers per call, and one unreadable slug fails the whole batch. Loop in threes and retry a failed batch one slug at a time.<\/p>\n
What This Costs You<\/h2>\n
Take the Rennes fixture and pull the best available price on each leg from the full 267-book board.<\/p>\n
\n\n\n\n\n\n\n\n
Leg<\/th>\nwinamax.fr<\/code><\/th>\nBest on the board<\/th>\nPosted by<\/th>\nDifference<\/th>\n<\/tr>\n<\/thead>\n
Rennes<\/td>\n2.10<\/td>\n2.20<\/td>\nbalkanbet.rs<\/code>, fanatics<\/code>, pointsbet.com.au<\/code><\/td>\n+4.8%<\/td>\n<\/tr>\n
Draw<\/td>\n3.95<\/td>\n4.04<\/td>\n1xbet<\/code><\/td>\n+2.3%<\/td>\n<\/tr>\n
Marseille<\/td>\n2.90<\/td>\n3.40<\/td>\nbetfair-ex<\/code> and four sharp exchanges<\/td>\n+17.2%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n
Across all 69 fixtures, the three Winamax slugs posted the best price on zero of 207 legs. The books that did keep turning up are wettarena<\/code> (42 legs), bet3000<\/code> (39) and the sharp exchange cluster (33 each). That is what line shopping across 350+ books<\/a> buys you over reading one national brand.<\/p>\n
One warning on that composite. The three best legs on this fixture sum to an implied 99.62%, which reads like free money and is not. Those prices come from four venues with four different liquidity profiles, and the exchange leg is a ladder level rather than a standing offer. Check limit<\/code> and exchangeMeta<\/code> before you believe a sub-100% composite, and treat it as a signal that your filter needs tightening.<\/p>\n
A Working Checklist<\/h2>\n