Compare
Fathom Analytics vs Umami
Umami is open source with a genuinely free, unlimited self-hosted option; Fathom is cloud-only, paid from day one, and asks nothing of you to run.
Side by side
| Fathom Analytics | Umami | |||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Vendor | Fathom Analytics (Conva Ventures Inc.) | Umami Software, Inc. | ||||||||||||||||||
| Pricing model | Subscription | Free tier + paid plans | ||||||||||||||||||
| Free tier | No | Yes | ||||||||||||||||||
| Deployment | Cloud | Cloud, Self-hosted | ||||||||||||||||||
| Open source | No | Yes (MIT) | ||||||||||||||||||
| Best for | Agencies and small businesses that want a no-maintenance, cookieless-compliant hosted analytics tool and are willing to pay a flat vendor subscription rather than self-host anything. | Developers and teams who want either a free, fully self-hosted analytics stack they control end-to-end, or a low-cost managed alternative billed by event volume instead of pageviews. | ||||||||||||||||||
| Pricing | Tiered monthly subscription billed on monthly pageview volume, with up to 50 sites included per tier; no free plan, 7-day free trial requires a card.
Prices read from the vendor's own page on September 21, 2026. Vendors change prices; check the source before you budget. | Self-hosting is free and unlimited; Umami Cloud uses usage-based pricing billed per event (pageview + custom events), with a free Hobby tier and paid Pro/Business/Enterprise tiers.
Prices read from the vendor's own page on September 21, 2026. Vendors change prices; check the source before you budget. | ||||||||||||||||||
| Features |
|
|
Verdict
Both are cookieless, privacy-first web analytics tools built to avoid a cookie consent banner, and both give a small business a simple, single-page dashboard instead of GA4's report sprawl. The real difference is who is responsible for hosting and operating the thing.
Fathom Analytics is cloud-only, closed source, and built by a bootstrapped company that hosts everything — there is no self-hosted option and never will be, by design. Pricing is tiered strictly by monthly pageviews, with up to 50 sites included at every tier, starting at $15/month with no free plan and no free trial without a credit card. In exchange, you get EU-hosted data routing for European traffic and zero operational responsibility.
Umami is MIT-licensed and genuinely open source: self-hosting is free and unlimited on your own infrastructure (you run the Node.js app and a Postgres/MySQL database), or you can use the managed Umami Cloud, which bills by total events — pageviews plus custom events — rather than by pageviews or site count, with a free Hobby tier covering one site and 100K events/month.
Choose Fathom Analytics if
- You want zero infrastructure to manage and are willing to pay a flat vendor subscription for that from day one.
- You run up to 50 sites and want them all covered under one plan without per-site pricing.
- EU-hosted data handling for European visitors specifically is a requirement.
Choose Umami if
- You want the option to self-host for free with no usage ceiling, and control your own infrastructure end-to-end.
- Your traffic mix includes a lot of custom events rather than just pageviews, where event-based Cloud billing may work out cheaper than a pageview-tiered competitor.
- You want a real free tier to start on, whether self-hosted or via Umami Cloud's Hobby plan.
What they share
Neither uses cookies or collects data that would trigger a consent-banner requirement, and both offer historical data import from Google Analytics to ease a migration. Both are genuinely aimed at the same buyer — someone who wants GA4's job done more simply and more privately — so the deciding question is almost entirely about hosting philosophy and budget rather than feature depth: pay someone else to run it and never think about infrastructure (Fathom), or take on that responsibility yourself in exchange for it costing nothing (Umami self-hosted). See how to choose a web analytics tool for how deployment model fits into the broader decision.
Last reviewed September 22, 2026