Umami is a privacy-first, open-source web and product analytics platform that gives site owners and product teams traffic, campaign, behavior, conversion and revenue reporting without cookies or surveillance, run either self-hosted or in the cloud.

What it is

Umami is a TypeScript analytics platform released under the MIT licence. It collects traffic, campaigns, behavior, conversions and revenue in one place, and it states plainly that it uses no cookies and performs no surveillance. It ships in two shapes: a self-hosted application that an operator runs against their own PostgreSQL database, and a cloud offering reachable from the project homepage at umami.is. The repository carries 38,878 stars and 8,051 forks, which places it among the more widely adopted projects in the Data & Analytics / Web & Product Analytics category, and it sits alongside the web-analytics, product-analytics, audience-segmentation, cohort-analysis and user-journey topics its maintainers tag it with.

The concrete problem it solves is ownership of analytics data. The project positions itself as an open alternative to Google Analytics and to Mixpanel. Both of those products require page instrumentation that reports visitor behavior to a third party, and Google Analytics in particular pushes sites toward cookie consent machinery. Umami replaces that arrangement with an application you deploy yourself, writing events into a PostgreSQL database you control. For teams that have been told to stop shipping visitor data to external vendors, or that simply want one place for traffic and conversion numbers without a consent banner attached to every page, the swap is the whole point of the project.

Key capabilities

• Cookie-free, surveillance-free measurement, so visitor tracking does not depend on client-side cookie storage.
• Reporting across five named areas in one interface: traffic, campaigns, behavior, conversions and revenue.
• Analysis features implied by the project topics: audience segmentation, cohort analysis, user journey views, charts and statistics.
• Environment configuration through a .env file, with DATABASE_URL as the required setting and the connection format postgresql://username:mypassword@localhost:5432/mydb.
• Optional API_URL to redirect internal UI API calls, either to a relative path such as /internal-api or to an absolute URL proxied through the local /api route, and BASE_PATH for serving under a subpath.
• Two-factor authentication, enabled by setting TWO_FACTOR_ENCRYPTION_KEY to a 64-character hex string generated with openssl rand -hex 32; the feature cannot be enabled or required until that key exists.
• Official Docker image at docker.umami.is/umami-software/umami:latest, plus a Docker compose file that runs Umami together with a PostgreSQL database.

Who uses it and how

• Self-hosters running the application on a server with Node.js 18.18 or later and PostgreSQL v12.14 or later, then placing it behind a reverse proxy such as nginx or changing the application port to serve it directly.
• Operators who prefer containers, deploying the published Docker image alongside PostgreSQL using the supplied compose file with docker compose up -d.
• Teams migrating off Google Analytics or Mixpanel that need the data to stay inside infrastructure they already run.
• Organizations that want a managed option instead of operating a database, using the cloud service at umami.is.
• Contributors and integrators drawn from a large base: 8,051 forks and 127 open issues indicate active outside participation, not a single-maintainer script.

Getting started

The full guide lives at umami.is/docs. The fastest path is Docker: pull docker.umami.is/umami-software/umami:latest and run docker compose up -d, which starts Umami with PostgreSQL. From source, clone the repository, run pnpm install, create .env with DATABASE_URL, then pnpm run build and pnpm run start, which serves on http://localhost:3000 by default.

How it compares

Against Google Analytics and Mixpanel, the differences are licence, hosting and data ownership rather than feature counts. Umami is MIT-licensed and can be self-hosted, so the event data lands in a PostgreSQL database the operator controls, while Google Analytics and Mixpanel are proprietary services that receive the data on their own infrastructure. Umami also offers a cloud option, so choosing it does not force self-hosting, but the self-hosted route is the one that removes third-party data transfer and per-event vendor cost entirely.

When to use it — and when not to

Pick Umami if you are willing to operate a PostgreSQL database, a Node.js 18.18 or later runtime and a reverse proxy, and if keeping analytics data in-house matters more than zero-ops convenience. Do not pick it if nobody on the team wants to run a database and a web server, since the self-hosted path assumes all three pieces. Note two practical cautions: the build step creates a login user named admin with the default password umami, which must be changed immediately, and the README covers installation well but defers nearly everything else to the external docs site.

git clone https://github.com/umami-software/umami.git
cd umami
pnpm install

DATABASE_URL=connection-url

postgresql://username:mypassword@localhost:5432/mydb

pnpm run build

pnpm run start

docker pull docker.umami.is/umami-software/umami:latest

docker compose up -d

git pull
pnpm install
pnpm build

docker compose pull
docker compose up --force-recreate -d