Everything comes back from a single call, which makes it cheap to wire up and hard to misuse — you ask what is being played and get scores, lineups, events and odds in one structured response. The flip side is scope: it is explicitly limited to live, upcoming and recently finished matches, so season tables and historical head-to-heads are not in reach.
A small MCP server over SoccerDataAPI, focused on matches that are live, about to start, or just finished. One call returns the global list of active matches with team names, kickoff time, stadium and current score, and the detail attached to each — goal breakdown by half, extra time and penalties, substitutions, cards, starting XI and bench with injury status and formation, win/draw/lose, over/under and handicap odds, and the league's name, country and competition format.
- Real-time information about football matches in progress worldwide, returned as structured data — `get_livescores`
- Match basics: home and away teams, kickoff time and date, stadium, current score, and whether the match is scheduled, in progress or finished
- Event detail: who scored, when and how, substitutions, yellow and red cards, penalties, and the goal breakdown across first half, second half, extra time and shootout
- Team lineups with the starting XI, the bench, injury status and formation
- Odds for win, draw and lose, over/under, and handicap markets, plus league name, country and competition format
A SoccerDataAPI account — sign up there and put the key in `AUTH_KEY`, in a `.env` file the client passes with `--env-file`. Python 3.12+ and the `uv` package manager; clone the repository, run `uv sync`, and launch `server.py` from `src/`, or install via Smithery. The published project is `mcp-soccer-data` 0.1.0. Scope is deliberate: live, upcoming and recently finished matches only — this is not a historical results archive.
One command plus a key — npx -y @smithery/cli install @yeonupark/mcp-soccer-data --client claude, then supply credentials