AcesAPI Review 2026

AcesAPI is a tennis data API for ATP and WTA singles, built set by set. It holds more than 124,000 finished matches back to 2007, with the game score of every set and the tiebreak points, plus tournament draws, player profiles, weekly rankings, head to head and an Elo rating for each player (checked November 2026).

You call it two ways. Every tool is one REST endpoint, POST /v1/{tool} on mcp.acesapi.com, and the same 14 tools run as an MCP server for Claude, Cursor and ChatGPT. The site is independent and says so: it is not affiliated with the ATP or the WTA.

AcesAPI covers one sport and goes deep on it. The data comes in five groups:

• Matches: singles results with every set's games and tiebreak score. Main draws start in 2007, qualifying in 2022.
• Tournaments and draws: dates, city, surface, a Grand Slam flag and each round's matches.
• Players: profiles, career records, head to head and records split by surface.
• Rankings: the weekly top 100 to 150 for both tours. The docs mark this group partial, because only 35 to 49 lists a year per tour are kept.
• Ratings: its own Elo, overall and on hard, clay and grass, before and after every match.

The scope is stated plainly. There is no doubles, no serve statistics and no live point-by-point feed. Data refreshes once a day, early morning US Eastern, so a match in progress shows the score at the last update. For live scores, pair it with a live tennis provider from the sports data API list.

Two things set AcesAPI apart. The first is the surface Elo. It computes the rating twice for every player, once from all matches and once from matches on the same surface only, so a clay specialist reads differently from a grass one. The site publishes how the number is counted. Surfaces come from its own tournament reference and stay null where it cannot place a tournament.

The second is a sportsbook archive. get_odds returns opening and closing moneylines for Grand Slam matches since mid-2020 and for ATP Masters and WTA 1000 matches since August 2024, with the games handicap and total where posted. Prices refresh hourly, only for matches in the next four days, and nothing is in play. That suits a model that wants a closing line next to the result and the Elo. For wide odds coverage, see the sports odds API guide.

Every account gets 500 free credits each month with no card. The free plan includes all 14 endpoints, REST and MCP. A call costs 1 to 5 credits: 1 for a lookup like search_players, 2 for results and rankings, 3 for head to head and surface splits, and 5 for get_player and get_odds. Empty results are not charged (checked November 2026).

Paid plans are monthly. Starter is $19 for 10,000 credits, Pro is $49 for 30,000 and Scale is $149 for 125,000. Wallet top-ups start at $5 for 2,000 credits and do not expire while the account is open.

The terms allow use inside your own app or model. Reselling the data as a feed is not allowed.

Sign up on acesapi.com/signup for the free plan: 500 credits a month, no card (checked November 2026). The docs say to verify your email, then create a key from the dashboard. The terms add that the site checks your email before it issues a key.

Send the key in the x-api-key header on every request. Authorization: Bearer works too, and the MCP endpoint also takes ?key=. Claude and ChatGPT connect to the MCP server with a sign-in instead of a key. The free plan is not a trial: the balance refills to 500 once a month while you have no paid subscription.

This sample calls search_players, a 1-credit tool, on https://mcp.acesapi.com with the key in x-api-key (from the docs, checked November 2026).

import os

import requests

resp = requests.post(
    "https://mcp.acesapi.com/v1/search_players",
    headers={"x-api-key": os.environ["ACESAPI_API_KEY"]},
    json={"name": "sinner"},
    timeout=10,
)
resp.raise_for_status()

body = resp.json()
for player in body.get("data", body)["players"]:
    print(player["player_id"], player["name"], player["country"], player["current_rank"])

It returns the players that match the name, each with a player_id, name, country and current_rank. The docs say REST wraps results in a data key, so the sample reads either shape. Pass the player_id to get_player for the profile and Elo ratings. There is also an official acesapi client on PyPI.

The rate limit is the same on every plan: 60 requests a minute, dropping to 10 a minute when your credit balance falls under 100 (checked November 2026). Credits are the real spend limit. A call costs 1 to 5 credits, an empty result is free, and a server error refunds the credits. A 402 means you are out of credits and a 429 means you hit the rate limit; send x-idempotency-key to make retries safe. Support is by email at support@acesapi.com, and paid plans add priority or dedicated support.

The docs cover each of the 14 tools with cURL and Python examples. A public OpenAPI 3.1 spec and a tool catalogue sit on mcp.acesapi.com (/openapi.json, /v1/tools). There is an official Python client, pip install acesapi, plus a status page and a data changelog. For assistants, see our sports MCP servers list.

• Tennis modellers and bettors who want surface-aware ratings, head to head and a closing line in one place.
• Analysts and researchers who need every set of both tours since 2007 as rows.
• Developers and AI builders who want an assistant to answer tennis questions over MCP with no client code.

Teams that need live tennis scores or serve statistics can run it beside a live feed. AcesAPI holds the history and the ratings, and the live feed holds the match in play.

AcesAPI is a focused tennis API with unusually clear coverage notes: it tells you what it holds, what is partial and what no public source publishes. The free plan is real, the credit prices are fixed and posted, and the surface Elo gives you something a plain results feed does not.

For historical ATP and WTA data, surface models and an AI assistant that talks tennis, it is a good first stop. Start with the free credits through acesapi.com, run get_ratings and get_head_to_head, and see whether the history fits your model.

Last updated: 5 November 2026 · How providers get listed