Plausible

Plausible

What is it?

Plausible is open source web analytics written in Elixir, pitched by its authors as a Google Analytics alternative built outside the ad-tech world. The tracking script sets no cookies and no persistent identifiers, and the server keeps no IP addresses or personal data; that is what the project's GDPR, CCPA and PECR compliance claim rests on. Plausible comes in two editions: a paid managed cloud running on EU-owned infrastructure, and a free, self-hosted Community Edition (CE) with a narrower feature set. Cloud subscriptions are the project's only source of funding.

Why is it interesting?

  • One-page dashboard: every key metric fits on a single screen, with no nested menus and no custom reports to build
  • Codeless tracking: outbound link clicks, form submissions, file downloads and 404 pages are counted without writing a single event
  • Single-page apps: pushState and hash-based routing work out of the box
  • Search Console integration: Google query data sits in the dashboard next to the traffic it brings in
  • APIs and export: an Events API to send events without the script, plus a Stats API and CSV export to get the numbers out
  • Google Analytics import: historical GA data comes along when you switch
  • MIT-licensed tracker: the server is AGPL-3.0, but the JavaScript snippet embedded in your pages is MIT, to keep AGPL virality away from the site that loads it

Use cases

  • Replacing Google Analytics on a public site while keeping sources, campaigns and entry pages, without cookies
  • Tracking conversion goals (sign-ups, contact clicks, downloads) through custom events
  • Publishing a public stats dashboard for an open source project
  • Reading traffic and Search Console queries in one place
  • Keeping raw event data in-house: CE lets you query the ClickHouse database directly

Good to know

Community Edition is not the cloud. The README lists features for both editions without telling them apart. CE ships without marketing funnels, ecommerce revenue goals, SSO or the Sites API. Its bot filtering relies on the User-Agent header and referrer-spam domains only, while the cloud also drops around 32,000 data-center IP ranges. CE is a long-term release published twice a year (latest: v3.2.1, May 15, 2026), and support is community-only.

Self-hosting means running two databases. The Docker Compose stack starts PostgreSQL for application data and ClickHouse for events. ClickHouse requires a CPU with SSE 4.2 or NEON, and the project recommends at least 2 GB of RAM. On a small VPS, Umami gets by with Node.js and PostgreSQL.

Two integration details. The install snippet is an inline script that calls plausible.init(): under a CSP without 'unsafe-inline', it has to move into its own file, and the instance's origin must be allowed in both script-src and connect-src. Outbound link tracking compares the link's host with the page's; a mailto: link has no host, so contact clicks only show up through a custom event.

SOURCES

REPOhttps://github.com/plausible/analytics
SITEhttps://plausible.io
LICENSEAGPL-3.0-or-later