// the find
rybbit-io/rybbit
🐸 Rybbit - open-source and privacy-friendly alternative to Google Analytics that is 10x more intuitive.
Rybbit is a self-hostable web analytics platform built on ClickHouse, targeting teams that want GA4-level depth (funnels, session replays, user journeys, retention) without Google's data collection. It runs as a Docker stack and tracks visitors without cookies. Aimed at product teams who've outgrown Plausible but won't touch GA4.
ClickHouse as the event store is the right call — columnar storage handles the aggregation queries analytics dashboards hammer constantly, and it scales without the index tricks you'd need on Postgres. The feature surface is genuinely competitive: custom events with arbitrary JSON properties, funnel analysis, session replay, and error tracking in a single self-hosted tool is unusual. The i18n coverage (12 languages in client/messages/) and a built-in API playground suggest this isn't just a weekend project. Cookieless fingerprinting means it works without a cookie banner in most jurisdictions, which is a real operational win.
AGPL-3 is a trap for anyone building a commercial product on top of this — you'd need to open-source your entire application or buy a commercial license; the README buries this. Self-hosting means running ClickHouse + Postgres (for auth/config) + the Next.js frontend + a backend API service, which is four things to keep alive and upgrade in lockstep — the operational surface is much larger than Plausible or Umami. Web Vitals and some features are cloud-only, which quietly undercuts the 'complete self-hosted alternative' pitch. Session replay storage at any real traffic volume is unaddressed — no documentation on retention policies, storage estimates, or how to prune old replays.