Umami is a privacy-first analytics platform. Traffic, campaigns, behavior, conversions, and revenue in one place — no cookies, no surveillance, self-hosted or in the cloud.
A detailed getting started guide can be found at umami.is/docs.
- A server with Node.js version 18.18+.
- A PostgreSQL database version v12.14+.
git clone https://github.com/umami-software/umami.git
cd umami
pnpm installCreate an .env file with the following:
DATABASE_URL=connection-url
NATS_URL=nats://localhost:4222Optional: set API_URL to change the base URL used by internal UI API calls.
Relative paths are served under BASE_PATH; absolute URLs are proxied through the local /api route.
For example, API_URL=/internal-api or API_URL=https://api.example.com/api.
Optional: set TWO_FACTOR_ENCRYPTION_KEY to a 64-character hex string to enable two-factor
authentication. Generate one with openssl rand -hex 32. Two-factor authentication is unavailable
and cannot be required until this key is set.
MCP is disabled by default. Set MCP_ENABLED=1 to enable the /mcp endpoint, then
authenticate with an API key generated under Settings → API keys.
The connection URL format:
postgresql://username:mypassword@localhost:5432/mydbThis fork keeps login sessions and its read-through cache in NATS JetStream KV, so every login can be revoked (logout deletes it). Umami refuses to start if NATS is not configured or a bucket cannot be read. It binds to existing buckets and never creates them.
| Variable | Required | Default | Description |
|---|---|---|---|
NATS_URL |
yes | — | NATS server URL(s), comma-separated |
NATS_CREDS |
no | — | Path to a .creds file (JWT + nkey) |
NATS_TOKEN_FILE |
no | — | Path to a token file sent as the password, re-read on every connect (a rotating Kubernetes ServiceAccount token for an auth callout); the user is NATS_USER, else SERVICE_IDENTITY. Used when NATS_CREDS is not set |
NATS_USER / NATS_PASSWORD |
no | — | User/password auth, used when neither NATS_CREDS nor NATS_TOKEN_FILE is set |
NATS_KV_AUTH_BUCKET |
no | umami_auth |
Bucket for login sessions |
NATS_KV_CACHE_BUCKET |
no | umami_cache |
Bucket for cached websites, links and pixels |
Session lifetimes: password and two-factor logins last 1 hour, SSO logins last OIDC_SESSION_DURATION,
and share-to-SSO tokens last 24 hours. The expiry is stored with each entry and checked on read. If the
auth bucket allows per-message TTLs (--marker-ttl), the server also removes sessions when they expire.
Give both buckets a max age (--ttl) at least as long as the longest session (24 hours unless OIDC_SESSION_DURATION is longer). The NATS user needs publish rights on
$KV.<bucket>.> and access to the $JS.API.STREAM.INFO, $JS.API.DIRECT.GET and $JS.API.STREAM.MSG.GET
APIs for KV_<bucket>.
Umami can authenticate users against any OpenID Connect provider using the authorization code flow with PKCE. Users are provisioned automatically on first login, and their umami role is derived from role claims in the ID token (or userinfo) on every login — the identity provider is the source of truth for access.
When OIDC is enabled, local password login is disabled (both the login form and the API) unless
OIDC_ALLOW_LOCAL_LOGIN is set.
| Variable | Required | Default | Description |
|---|---|---|---|
OIDC_ISSUER |
yes | — | Issuer URL; discovery document is read from <issuer>/.well-known/openid-configuration |
OIDC_CLIENT_ID |
yes | — | OAuth client ID |
OIDC_CLIENT_SECRET |
no | — | Omit for a public client using PKCE; set for a confidential client |
APP_URL |
yes | — | Public base URL of your umami instance, e.g. https://analytics.example.com |
OIDC_ROLE_ADMIN |
* | — | Comma-separated provider role keys mapped to the umami admin role |
OIDC_ROLE_USER |
* | — | Comma-separated provider role keys mapped to the umami user role |
OIDC_ROLE_VIEW_ONLY |
* | — | Comma-separated provider role keys mapped to the umami view-only role |
OIDC_SCOPES |
no | openid profile email |
Requested scopes; add any provider-specific scope needed to receive role claims |
OIDC_ROLES_CLAIM |
no | roles |
Claim containing the user's roles (object map, array, or delimited string); exact names — including URN/URL-style names — are matched first, and dot notation reads nested claims (e.g. realm_access.roles) |
OIDC_ALLOW_LOCAL_LOGIN |
no | off | Keep username/password login available alongside SSO (break-glass) |
OIDC_PROVIDER_NAME |
no | SSO |
Name shown on the login button |
OIDC_RP_LOGOUT |
no | off | Also end the identity provider session on logout |
OIDC_AUTO_LINK_USERNAME |
no | off | Link an SSO login to an existing local account with the same username (only from preferred_username, or from email when email_verified is true) |
OIDC_SESSION_DURATION |
no | 24h |
Lifetime of umami sessions created via SSO (e.g. 12h, 30m, 7d) |
* At least one of the three role mappings must be set. A user with none of the mapped roles is denied access.
Do not combine DISABLE_LOGIN with OIDC — the login page must render to show the single sign-on button.
Users who were already logged in with a local account keep their existing sessions when OIDC is enabled;
rotate APP_SECRET to force everyone through the identity provider.
Provider setup: register a web application using the authorization code flow with PKCE (no client
secret needed; set OIDC_CLIENT_SECRET if you use a confidential client instead), set the redirect URI to
<APP_URL>/api/auth/oidc/callback and the post-logout redirect URI to <APP_URL>/login, define roles
(e.g. umami-admin, umami-user) and grant them to your users. Make sure the provider includes the roles
in the ID token or userinfo response — the claim name and any extra scope this requires vary by provider,
so set OIDC_ROLES_CLAIM and OIDC_SCOPES accordingly.
pnpm run buildThe build step will create tables in your database if you are installing for the first time. It will also create a login user with username admin and password umami.
pnpm run startBy default, this will launch the application on http://localhost:3000. You will need to either proxy requests from your web server or change the port to serve the application directly.
Umami provides Docker images as well as a Docker compose file for easy deployment.
Docker image:
docker pull docker.umami.is/umami-software/umami:latestDocker compose (Runs Umami with a PostgreSQL database):
docker compose up -dTo get the latest features, simply do a pull, install any new dependencies, and rebuild:
git pull
pnpm install
pnpm buildTo update the Docker image, simply pull the new images and rebuild:
docker compose pull
docker compose up --force-recreate -d