xeve developer documentation
xeve records what one person actually does — applications and websites used, coding time by project and language, GitHub activity, listening history, health samples, heart rate, locations — and this is how you get it back out.
Everything here reads your own data. There is no endpoint that takes a user id, and none can be added: the server derives the user from the credential and Postgres row-level security filters every query.
Start here
| API reference | Base URL, authentication, every endpoint, error shapes |
| OAuth 2.0 | The authorization flow, scopes, dynamic client registration |
| MCP server | Connect Claude, ChatGPT or any MCP client to your data |
| For AI agents | When to reach for xeve, and how to call it |
Machine-readable
| File | What it is |
|---|---|
| https://xeve.io/openapi.json | OpenAPI 3.1 description of the API, including security schemes and scopes |
| https://xeve.io/.well-known/oauth-authorization-server | OAuth 2.0 authorization server metadata (RFC 8414) |
| https://api.xeve.io/.well-known/oauth-protected-resource | Protected resource metadata (RFC 9728) |
| https://xeve.io/.well-known/mcp | MCP server manifest: endpoint, transport, auth, capabilities |
| https://xeve.io/llms.txt | What xeve is, when to use it, and every key URL |
| https://xeve.io/sitemap.xml | Every public page |
Every page on xeve.io also answers Accept: text/markdown with a Markdown
representation of itself, and serves the same document at its .md sibling —
https://xeve.io/docs/api.md.
Self-hosting
The backend is open source and runs in two containers, app and Postgres. See
the repository and docs/self-hosting.md.
Every URL in these docs is https://api.xeve.io for the hosted service and
your own origin for a self-hosted one.