OURA_CLIENT_ID — Client ID of your Oura application from developer.ouraring.com
OURA_CLIENT_SECRET — Client secret of your Oura application
OURA_REDIRECT_URI — OAuth redirect URI; must match the Oura application byte for byte
OURA_API_MODE — sandbox for Oura's synthetic data (no auth needed), production for your own
OURA_TZ — IANA timezone deciding what counts as today. Set it explicitly on servers, whose clocks run in UTC
OURA_TOKEN_STORE — Where the OAuth flow stores tokens. Written by the server, not set by hand
OURA_PUBLIC_URL — Public https address, without a trailing slash. Enables OAuth 2.1 for Claude.ai and ChatGPT; leave empty for local or Bearer-token clients
OURA_OAUTH_ALLOWED_REDIRECT_ORIGINS — Comma-separated trusted OAuth redirect origins; defaults to Claude.ai and ChatGPT
Claude Code has no install link for a local command. Copy the command below and run it in a terminal.
New to MCP? Model Context Protocol lets AI clients reach external data.
Install this server, connect it once, then ask about your sleep in plain
language. No coding involved.
Works locally in Claude Code and Codex, and remotely in claude.ai and ChatGPT
once it is deployed to a server of your own.
Compact by default. Tools return per-day values plus statistics with a
trend, not the raw API payload. Raw responses stay one raw=True away.
A month of heart-rate data shrinks by more than 10×.
One codebase, two transports.stdio for local use, streamable-http
for remote. A flag apart, not a rewrite.
Timezone-correct. Oura filters some endpoints by an internal UTC
timestamp while returning a local day field, and returns heart-rate
timestamps in UTC. Both quietly lose or misplace data outside UTC. This
server handles it — see Timezone handling.
Survives flaky networks. Follows next_token pagination, retries
dropped connections with exponential backoff, and turns HTTP status codes
into messages that say what to fix.
Try before authorizing. Oura's sandbox works with no credentials at all.
Check that data flows (hits Oura's sandbox, no auth required):
uv run python -m my_oura_mcp.smoke
Connect it locally — the server prints the commands with the absolute path
already filled in, and writes nothing itself:
uv run my-oura-mcp install
Then ask your client for an Oura summary. The get_status tool reports which
mode the server is in.
Tools
Tool
Returns
Default range
get_daily_summary
Sleep, readiness and activity scores at once
7 days
get_sleep
Sleep stages, efficiency, HRV, resting HR, breathing, temperature
7 days
get_sleep_score
Daily sleep score only — lighter than get_sleep
7 days
get_readiness
Readiness score, HRV balance, temperature deviation
7 days
get_activity
Activity score, steps, calories
7 days
get_heartrate
Per-minute heart rate collapsed to daily stats
3 days
get_spo2
Blood oxygen during sleep, breathing disturbance index
7 days
get_stress
Time under load and in recovery
7 days
get_heart_health
Cardiovascular age, VO₂max
30 days
get_tags
Tags you entered in the Oura app
30 days
get_status
Server mode and authorization state
—
Every data tool takes either days_back or an explicit start_date/end_date
pair (YYYY-MM-DD), plus raw to get Oura's untouched response.
Using your own data
The sandbox returns synthetic data. For your own you need an Oura application
and a one-time authorization — no review to pass, a fresh application works
immediately.
Put OURA_CLIENT_ID and OURA_CLIENT_SECRET into .env
3
uv run my-oura-mcp auth — opens a browser, stores tokens with mode 600
4
Set OURA_API_MODE=production in .env
When registering: Redirect URI is http://localhost:8765/callback, matched
byte for byte. Scopes are daily, heartrate, tag, spo2, stress,
heart_health. Everything else is arbitrary.
Tokens refresh themselves from there. Check with my-oura-mcp auth --status,
forget them with auth --logout.
Personal Access Tokens no longer work: Oura stopped issuing them in December
2025.
Refresh tokens are single-use. Each refresh kills the old one, so two
instances sharing a token store knock each other out. The symptom is a 400
mentioning single use; the cure is re-running auth and keeping one instance.
Configuration
Everything lives in .env (see .env.example). Secrets never reach git.
Variable
Purpose
OURA_CLIENT_ID / OURA_CLIENT_SECRET
Oura application credentials
OURA_REDIRECT_URI
Must match the application exactly
OURA_API_MODE
sandbox (synthetic data) or production
OURA_TZ
Timezone deciding what "today" means. Set explicitly on servers
OURA_MCP_TOKEN
Shared secret guarding the HTTP endpoint; also the consent-page password
OURA_PUBLIC_URL
Public https address. When set, enables OAuth for Claude.ai and ChatGPT
OURA_OAUTH_ALLOWED_REDIRECT_ORIGINS
Comma-separated OAuth client origins; defaults to Claude.ai and ChatGPT
OURA_TOKEN_STORE
Where the OAuth flow writes tokens. Not set by hand
OURA_CACHE_DB
SQLite cache file. An empty value disables caching
Running it
One codebase, two transports: stdio next to a client on this machine, HTTP on
a server so the web clients and your phone can reach it.
Locally
Client
Command
Claude Code
claude mcp add --scope user oura -- uv --directory PATH run my-oura-mcp
Codex
codex mcp add oura -- uv --directory PATH run my-oura-mcp
uv run my-oura-mcp install prints these with the absolute path already filled
in. --scope user makes the server visible from any directory; without it, only
from where the command ran. Verify with claude mcp list / codex mcp list.
Nothing is exposed and no network is involved. To debug the transport itself:
uv run my-oura-mcp --transport http --port 8000 — no secret needed on loopback.
On a server
This is what makes the server reachable from any device, from Claude.ai and from
ChatGPT. Step by step in docs/DEPLOY.md (Russian): how to
let traffic in, how to move Oura authorization across, how to connect web
clients.
Client
How it connects
Claude.ai, ChatGPT
OAuth. No client ID or secret to enter — the client registers itself and the consent page asks for OURA_MCP_TOKEN
Claude Code
claude mcp add --scope user --transport http oura URL --header "Authorization: Bearer TOKEN"
OURA_PUBLIC_URL turns OAuth on: the server becomes its own authorization
server with dynamic client registration, restricted to https://claude.ai and
https://chatgpt.com. Another client needs its exact origin in
OURA_OAUTH_ALLOWED_REDIRECT_ORIGINS; the consent page names the client and its
return origin before asking for the secret.
Two ways to expose it, and the choice is not cosmetic. Caddy is simpler, but its
certificate lands in Certificate Transparency — a public log revealing that this
address hosts a service. A Cloudflare Tunnel opens no inbound ports at all.
If a VPN lives on the same host, only the tunnel will do.
Which one
stdio, local
HTTP, on a server
Claude Code and Codex on this machine
yes
yes
Other devices
no
yes
Claude.ai, ChatGPT, phone
no
yes, over OAuth
Needs a domain and a host
no
yes
Data leaves this machine
no
yes, to your server
Keep exactly one live instance: Oura's refresh token is single-use, and two
servers sharing a token store will knock each other out of authorization.
Timezone handling
Three separate bugs came from Oura's date semantics, all of which lost data
silently rather than raising an error. Worth knowing if you build against this
API yourself:
sleep and daily_activity are filtered by an internal UTC timestamp,
not by the day field Oura itself returns. At UTC+3 a night that starts
after midnight lands in the previous UTC day: asking for 28..28 returns
nothing while the record with day=28 plainly exists. The server widens the
window and trims by day afterwards. Verified by sweeping every endpoint;
the other six behave.
heartrate returns timestamps in UTC. Grouping by the first ten
characters of that string splits a local day in two, pushing 00:00–03:00
local into the previous day — exactly the resting heart rate you care about.
Grouping uses OURA_TZ.
Oura returns several sleep records per day — the night plus naps. Picking
an arbitrary one lets a 12-minute nap displace a full night. The record typed
long_sleep wins, or the longest one; naps are reported separately as
naps_h so their HRV never averages with the night's.
Security
.env, the token store and the cache are in .gitignore. Verify before
committing: git status --porcelain.
The token store and SQLite cache are written with owner-only (600) file
permissions.
The HTTP endpoint is guarded by OURA_MCP_TOKEN using a constant-time
comparison. The access model is deliberately simple: one secret, one owner,
no per-user separation.
The server refuses to start on a non-loopback address without a secret
rather than quietly serving health data to the open internet. Try it:
uv run my-oura-mcp --transport http --host 0.0.0.0.
/healthz is intentionally open — a reverse proxy needs it, and it returns
nothing but ok.
Caddy strips the Authorization header from its logs.
Caching
Older days go into SQLite. The two most recent completed days are rechecked on
every request because late syncs can update them; older history is served from
the cache without a network call.
uv run my-oura-mcp cache --status # what is cached
uv run my-oura-mcp cache --clear # forget it
Three things worth knowing. Today is never cached — Oura is still writing
it. Empty days are not cached either: an empty day means either "did not
wear the ring" or "has not synced yet", and the second resolves itself within
hours, whereas a cached blank would last forever. The mode is part of the
key, so sandbox data cannot surface in production.
Per-minute heart rate bypasses the cache: its rows carry no day field.
MCP resources
Clients that support resources can read oura://today, oura://yesterday,
and oura://week. They provide the same sleep/readiness/activity summaries
as the tools, without choosing arguments manually.
Skill with recipes
skills/oura ships a Claude-oriented skill — not more tools, but workflows
on top of them: whether sleep is actually improving, whether today can take
load, what the body was doing on a bad day, whether a change in routine did
anything. Each is a sequence of calls plus a way to reason about the answer,
which no single tool can express.
Install it by copying into your client's skills directory:
cp -r skills/oura ~/.claude/skills/
The recipes are checked against the code by tests: a field name that no tool
returns fails uv run pytest instead of quietly sending the model nowhere.
Codex can also load these repository skills after copying them to its skills
directory:
oura interprets data through the MCP server. oura-mcp-maintenance guides
safe changes to this repository and its Claude/ChatGPT/Codex integration.
When something doesn't work
Claude says there are no Oura tools. The server didn't connect. claude mcp list shows its state. A common cause is a relative path where a full one is
required — uv run my-oura-mcp install prints the command with the right one.
"Авторизация не пройдена — токенов нет". The server is in production mode
but has never signed in to Oura. Run uv run my-oura-mcp auth; check token state
with uv run my-oura-mcp auth --status.
401 against a server on a VPS.OURA_MCP_TOKEN doesn't match. Header
values are sent verbatim, so the word Bearer and the space belong to the value:
Bearer abc123, not abc123.
Data comes back for the wrong day.OURA_TZ isn't set. A server clock is
almost always UTC, so "today" starts hours off from yours and a night's sleep
lands in the previous day. Set it explicitly, e.g. OURA_TZ=Europe/Moscow.
"refresh-токен отвергнут". Oura's refresh token is single-use, and this
happens when a second instance spent it. Keep exactly one alive: once the VPS is
up, point local Claude Code at it too. Recover with my-oura-mcp auth.
Requests to api.ouraring.com fail with SSL_ERROR_SYSCALL or a timeout.
Usually not the server: the client retries four times with backoff. If that
doesn't help, a VPN generally does.
claude.ai won't connect to your server. Check that OURA_PUBLIC_URL is set
and matches the connector URL character for character, including https:// and
no trailing slash. The startup banner says whether OAuth came up. Beyond that,
see docs/DEPLOY.md.
You press Allow on the consent page and nothing happens. Check the server
logs: a POST /oauth/consent returning 303 with no POST /token after it
means the browser blocked the hop back to claude.ai. That is what an over-strict
Content-Security-Policy looks like — and curl cannot reproduce it, since it
does not enforce CSP at all.
"Запрос устарел" / request expired. Authorization requests live in process
memory, so restarting the server invalidates any consent page already open. The
secret is not the problem — go back to claude.ai and start the connection again.
Development
uv run pytest
Tests never touch the network; Oura's responses are stubbed with respx.
Tool tests go through mcp.call_tool() rather than calling the functions
directly — some bugs only appear on the real protocol layer, where MCP clients
pass declared defaults as explicit arguments.
For Codex and other coding agents, repository instructions and a task/review
structure live in AGENTS.md and docs/agents.
x86_64 macOS:cryptography 49+ ships no wheel for this platform and tries
to build from Rust sources. pyproject.toml pins 48.0.0 for it specifically;
Linux and native arm64 are untouched. This bites Apple Silicon too whenever
Homebrew lives in /usr/local rather than /opt/homebrew — check with
file $(which python3).
Flaky network: if requests to api.ouraring.com fail with
SSL_ERROR_SYSCALL or time out, it usually isn't the server. The client makes
four attempts with backoff; beyond that, try a VPN.
See docs/ROADMAP.md for what's planned and what was
deliberately deferred.