Steve Gomori's training data, as analytical tools
Every run Steve Gomori records is parsed from its raw Garmin file, reduced to the metrics that describe an endurance athlete's training, and exposed through the Model Context Protocol. Any MCP-compatible AI client can connect and ask about it.
The tools return shaped aggregations rather than raw data — volume and load over a period, intensity distribution, pace progression against grade, race readiness — and stop short of drawing conclusions. Interpretation is left to whichever client is asking.
Connect your own client
The server speaks Streamable HTTP and authenticates with a bearer token. Point any MCP client at the endpoint below and it will discover the available tools on connect.
https://training.stevegomori.ca/mcpKeys are issued on request and are read-only — a client can query the training history and nothing else. Running your own instance needs no key from anyone: the project is self-hostable end to end, against your own Garmin exports.