Connect your AI assistant
Connect ChatGPT, Claude, or another OAuth-capable MCP client to RideIndex. With read access, your assistant can inspect your rides, routes, workouts, calendar, and fitness metrics. If you separately approve write access, it can create or modify workouts and scheduled sessions when you explicitly request it.
https://rideindex.app/api/mcp- 1In your assistant, add a custom connector and paste the URL above.Not seeing the option? Custom connector support varies between assistants and plans.
- 2It sends you to RideIndex to sign in and approve access. Choose read-only, or allow writes so it can build and schedule workouts too. No token to copy — the connector registers itself.
- 3Start a chat and ask it something. It reads your data and answers.
- “Explain what changed over my last six weeks”
- “Draft my 2x20 tempo workout for me to review”
- “Compare my attempts on this saved route”
- “Analyze the fade in my last long ride”
- “How's my form trending into next month?”
Using a different client?
Any OAuth-capable MCP client connects the same way — paste the URL, sign in, approve.
stdio-only clients (or a local desktop config) can bridge with mcp-remote, which opens a browser for the same sign-in:npx mcp-remote https://rideindex.app/api/mcp
A client that can't do an OAuth handshake at all (an SDK-built agent, say) can use a personal access token as a static Authorization: Bearer … header instead.
Why it works this way
We don't run a model over your rides. You connect an assistant you already have, it reads your data on your instruction, and you can disconnect it at any time from your settings.
When you ask your connected assistant to use RideIndex, RideIndex returns the data authorized by your connection directly to that client. The assistant provider handles that data under its own terms and privacy policy — RideIndex does not choose or operate the model for you.