Bikepackia for developers and AI agents

Last verified September 12, 2026

Bikepackia runs an MCP server that lets AI agents and assistants plan, refine and export bikepacking routes for a rider. An agent sends a plain-language brief, and Bikepackia returns a route built on real map data, with day stages and a GPX file. Agents sign in with a Bikepackia API key and spend that account's credits.

What you can build

• An AI assistant that plans a multi-day bike trip from a sentence like "3-day gravel loop from Lyon, avoid highways".
• A travel or outdoor app that hands riders a rideable route with day stages and a GPX file.
• A chat workflow that refines a route over several turns ("add a rest day before the climb") and reads back its safety scorecard.

How it works

The agent never draws the route itself. Bikepackia's planner turns the brief into routing requirements (start, end, days, surface, daily distance), and a routing engine computes the actual path over OpenStreetMap data. That's why every route an agent gets back follows real roads and tracks.

• The MCP server exposes four tools: planRoute, refineRoute, exportGpx and getRouteSafety. See the MCP server page.
• Each call is authenticated with an API key tied to a Bikepackia account. See the API keys page.
• Every planned route is saved to that account, so an agent can refine or export it minutes or hours later.

Pricing

Agent calls draw from the same credit balance the account uses in the Bikepackia apps. Planning and refining spend credits because they run the AI planner. Exporting GPX and reading the safety scorecard only read a route that's already saved, so they don't use the AI planner and don't spend credits.