Umami

Umami is a web and product analytics platform from Umami Software, Inc. It brings traffic, campaigns, behavior, conversions, retention, and revenue reporting into one dashboard without cookies or automatic collection of personal data. It can be self-hosted from the MIT-licensed source or used through Umami Cloud, and the latest release listed on GitHub is v3.4.0.

What It Is

Umami is a privacy-focused analytics tool for websites and products. A site adds a tracking script, and Umami reports pageviews, visitors, bounce rate, visit time, locations, devices, browsers, and referrers. Its homepage says it stores nothing personally identifiable, which it states supports GDPR, CCPA, and PECR obligations, and that most sites can drop the cookie banner.

Analytics Workflow

Umami covers several stages of analysis:

• Acquisition: channels, referrers, and automatic UTM tracking, with attribution and side-by-side comparison of two time periods.
• Behavior: custom events such as clicks, form submissions, signups, and downloads, plus Journey reports, heatmaps, and session replay.
• Conversion and retention: funnels built from pages and events, goals, retention reports grouped by first-visit day, and saved cohorts that can be applied as filters.
• Revenue: events tied to dollar amounts, so pages, campaigns, and segments can be compared by value.

MCP Access for AI Assistants

Umami has an official MCP server that connects Claude, ChatGPT, Cursor, and other MCP clients to analytics data. Its read-only tools cover traffic, events, sessions, funnels, goals, journeys, retention, attribution, revenue, and Core Web Vitals, and they respect existing website and team permissions. According to the homepage, it is available on Umami Cloud Pro plans and above and on self-hosted Umami v3.4 or later. In self-hosted installs, MCP is disabled by default and is enabled with an environment variable plus an API key.

Self-Hosting Setup

Installing from source requires Node.js 18.18+ and PostgreSQL v12.14+. Users clone the repository, install with pnpm, set a DATABASE_URL, build, and start the app on port 3000. The build creates database tables and a default admin login. Docker images and a Docker compose file that runs Umami with PostgreSQL are also provided. Optional two-factor authentication is available once an encryption key is set.