Nice — let's get you set up.
sportsdata is free, MIT licensed, and there's nothing to activate: no account, no licence key, no trial. Grab the build for your platform, then three steps and your AI client is talking to live sports data.
Opens the GitHub releases page — pick the asset matching your OS.
Install it
-
Unzip and run setup
Open a terminal where you unzipped it, then run setup once. It writes the MCP config for the clients it finds on your machine.
./sportsdata-mcp setup
macOS: the build is unsigned for now, so double-clicking will be blocked. Right-click the app → Open → Open, once, and macOS remembers it. -
Restart your AI client
Quit and reopen Claude Desktop, Cursor, or whichever MCP client you use — it reads the config at startup. You should see the sportsdata tools appear in the tool list.
-
Ask it something real
A good first test, because it hits several providers at once:
What's on in the AFL this round, and which book has the best price on each favourite?
Prefer no download?
The server is on PyPI, so one line wires it into any MCP client — no app, no Python project, nothing to unzip.
uvx sportsdata-mcp serve
If something breaks
Open an issue on GitHub or email danieltomaro@icloud.com. Some bookmaker feeds are geo-restricted outside their home region — if a provider returns nothing, that's usually why.
Free, and staying that way
sportsdata is built and maintained by one person, and every feature is free for everyone whether you chip in or not — nothing is gated. If it saves you time, a coffee helps keep the feeds maintained.