Read-only OAuth · local tokens · agent-native health context

Your body already sends signals. Your agent should know how to listen.

Oura MCP is an unofficial, open-source bridge that gives AI agents safe access to your Oura readiness, sleep, activity, heart-rate, HRV, SpO2 and workout context through the official Oura Cloud API v2.

npx -y oura-mcp-unofficial setup
20tools
Sleep7h 10mstage-ready context
HRV48 mstrend, not diagnosis
Activity9,000steps + active minutes
Official APIOura Cloud API v2 OAuth2
Local-firsttokens under ~/.oura-mcp
20 toolsreadiness, sleep, activity, heart, SpO2
Agent-readyClaude, Cursor, Hermes, OpenClaw

For humans

No API archaeology. No token pasted into agent configs.

Oura developer setup is the hard part. This page gives the exact callback, scopes and commands so a non-technical user can delegate the install safely.

01

Create your Oura app

Open the Oura developer console and add this callback URL.

http://127.0.0.1:3000/callback
Open Oura apps
03

Run setup

npx -y oura-mcp-unofficial setup
npx -y oura-mcp-unofficial auth
npx -y oura-mcp-unofficial doctor

The toolbelt

Processed health signals, shaped for agents.

Activitydaily activity score, steps, calories and distance
Sleeplogs, duration, efficiency and stages when Oura provides them
Heartheart-rate records when account and scopes permit
Recovery contextHRV, readiness contributors and SpO2 availability by account/device
Workoutsworkout summaries, sessions and tags when granted

For agents

A manifest that tells agents how not to break things.

Agents should call oura_agent_manifest and oura_connection_status first. Hermes gets direct tool names and explicit reload guidance so it does not restart the gateway for normal Oura access.

Install the unofficial Oura MCP server for me.
Repo: github.com/davidmosiah/ouramcp
Scopes: daily heartrate personal sleep workout spo2
Then run setup, auth and doctor.
Keep tokens local. No medical diagnosis.

Trust boundary

Health data is not a toy payload.

This MCP exposes processed Oura Cloud API v2 data. It does not expose raw accelerometer telemetry, private Google endpoints, or clinical diagnosis. Raw JSON is explicit; summary and structured modes are safer defaults.

summaryminimal safe fields
structurednormalized agent data
rawexplicit upstream JSON
0600local token permissions

Open source quantified-self infrastructure

Make personal health agents safer, clearer and more useful.