FighterAPI Review 2026

FighterAPI is a data API for mixed martial arts. It holds every UFC card since 2000, with the result, method, round, time and judges' scores for each fight, plus a stat sheet for nearly every one. PFL (from 2012) and Bellator (2009 to 2024) sit beside it (checked November 2026).

You reach it two ways. Each of the 11 tools is a REST call, POST /v1/{tool} on mcp.fighterapi.com, and the same tools run as a hosted MCP server for Claude, ChatGPT and Cursor. The site says plainly that it is independent and holds no league rights.

The docs split the data into six groups, and the data atlas marks how complete each one is:

• Cards and results: events, fights in card order, winner, method, round, time, judges' scores and referee, from 2000.
• Fight stat sheets: strikes by target and position, takedowns, control time, guard passes, reversals and submission attempts. A few dozen fights, mostly 2016 to 2019, have no sheet.
• Round by round: the same measures for each round, checked against the fight totals.
• Fighters: birth date, height, reach, stance, gym, record by method and career rates.
• Ratings: an Elo for every UFC fighter, before and after each fight.
• Odds: opening, closing and current moneylines and the total-rounds line. The atlas marks this group partial.

Data refreshes once a day at 10:00 UTC, so a Saturday card appears on Sunday morning. There is no live feed and no play-by-play. For live scores, add a live provider from the sports data API list.

Three things set FighterAPI apart. First, the round-by-round lines. Most free sources stop at fight totals, and this one splits them by round and flags a fight whose rounds do not add up. That lets you measure fatigue or a fast start.

Second, the Elo. FighterAPI computes it from every UFC result since 2000, starting each fighter at 1,500, and recomputes it daily. get_ratings filters by weight class.

Third, the odds. get_odds returns opening, closing and current moneylines by sportsbook for UFC fights since June 2020, and get_line_movement returns the snapshots in between. Lines are pre-game only, with a median of 5 to 7 books per fight. Early PFL and Bellator odds come from fewer books. For broader pricing, see the historical odds guide.

Every account gets 500 free credits each month with no card, on all 11 endpoints. A call costs 1 to 10 credits: 1 for list_events and search_fighters, 2 for get_event, get_schedule and get_ratings, 5 for get_fight, get_fighter, get_fighter_fights, get_leaders and a single-fight get_odds, and 10 for get_line_movement (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 let you store the data, train models and build commercial products that show users what you derive. You may not pass the data on as a feed or API of its own.

Create a free account at fighterapi.com/signup. You get 500 credits every month with no card, and the free plan covers all 11 endpoints over REST and MCP (checked November 2026). Verify your email, then create a key in the dashboard.

Send the key in the x-api-key header. Authorization: Bearer works too, and the MCP URL takes ?key=. Claude and ChatGPT skip the key: you add the MCP server as a connector and sign in once. The free plan is not a trial. The balance tops back up to 500 each month while you have no paid plan.

This sample calls search_fighters, the one-credit lookup that turns a name into a fighter id. The fighter endpoints take that id. It posts JSON to https://mcp.fighterapi.com/v1/search_fighters with the x-api-key header.

import os
import requests

resp = requests.post(
    "https://mcp.fighterapi.com/v1/search_fighters",
    headers={"x-api-key": os.environ["FIGHTERAPI_API_KEY"]},
    json={"query": "topuria"},
    timeout=30,
)
resp.raise_for_status()
body = resp.json()
for f in body.get("data", body)["fighters"]:
    print(f["fighter_id"], f["name"], f["weight_class"], f["pro_record"])

Each match comes back with fighter_id, name, weight_class and pro_record. The call follows the docs example (from the docs, checked November 2026). The docs show the result bare, while the OpenAPI spec wraps it in data, so the loop handles both. The official fighterapi package on PyPI wraps the same calls.

The limit is 60 requests a minute on every plan. It drops to 10 a minute while your balance is under 100 credits (checked November 2026). Credits are the real spend cap. A 429 means you sent too fast, a 402 means the balance is empty, and a 500 refunds its credits. Send x-idempotency-key to retry safely, and read X-Credits-Remaining on each response.

The docs have a page for each endpoint. The OpenAPI 3.1 spec is at mcp.fighterapi.com/openapi.json, and llms.txt lists the tools for agents. The official Python client is fighterapi on PyPI. A public status page shows probe results every five minutes, and a data changelog shows when each dataset last refreshed. Support is support@fighterapi.com.

Bettors and modellers get the closest fit. Closing lines sit next to the result, both stat sheets and both ratings, which is the table a UFC model needs. Analysts get per-round strike and takedown data without writing a scraper. Developers building AI tools get a hosted MCP server and 16 free Claude skills from the provider.

The scope is one sport and the past. For live in-fight data, or licensed feeds with league rights, pair it with a provider such as SportsDataIO or Sportradar. For odds across many sports, use the sports odds API guide.

FighterAPI is a focused, well-documented MMA data API with a real free plan. Round-level stats, an open Elo and a six-year moneyline archive in one place is rare, and the credit price list is easy to budget. Start with the free credits, pull one card end to end, and check the odds history against your model.

Last updated: 5 November 2026 · How providers get listed