Analytics cookies only (Google Analytics, PostHog). No ad networks, no cross-site tracking. See our privacy policy.
upstream: github.com/matomo-org/matomo (opens in new tab) · github.com/umami-software/umami (opens in new tab)
Pick Umami unless you genuinely need Google Analytics-grade reporting — it runs on half the RAM, installs in minutes, and is cookieless by default. Pick Matomo when you are migrating off GA and need e-commerce tracking, goals, and raw data you own, and accept the heavier PHP + MariaDB stack that comes with it.
Matomo and Umami sit at opposite ends of the self-hosted analytics spectrum. Matomo is the full-catalog Google Analytics replacement — goals, e-commerce tracking, raw data you own — while Umami is the featherweight dashboard that answers "how is my site doing" and gets out of the way. This is the rare comparison where you can rule one out in a minute: decide how much analytics product you actually need, and the winner falls out of that answer.
Umami is about as small as real analytics gets. It is a Next.js app backed by a single PostgreSQL database, licensed MIT, and it runs comfortably on a 1 GB box at a friendly 2 out of 5 difficulty. One script tag on your site, one clean single-page dashboard: pageviews, unique visitors, referrers, top pages, countries, devices, plus custom events. That is the whole product, and for most personal sites and small projects it is the whole requirement.
Matomo is the opposite bet. A PHP application with a MariaDB database behind it, licensed GPL-3.0, wanting 2 GB of RAM and rating a 3 out of 5 to deploy. In exchange it is the closest thing to a like-for-like GA replacement you can self-host: goals and e-commerce tracking built in, segmentation, campaign reports, and raw visitor-level data that is fully yours to export and query. Funnels and heatmaps are available too, though as paid premium plugins rather than free core features.
If you run a store, manage marketing campaigns, or your job involves proving which channel converts, Matomo is playing a game Umami does not enter. E-commerce reports, goal tracking, and segments are the workflow GA people expect, and Matomo is the self-hosted tool that preserves it. The raw data layer matters here too: when a question isn't answered by a built-in report, you can export the underlying visits and answer it yourself.
Umami v3's MIT build covers traffic, sources, pages, geography, devices and custom events, plus funnels, journeys, retention, cohorts, session replays, heatmaps and Web Vitals, several of which Matomo sells as premium plugins. What it lacks is e-commerce reporting and a raw visitor-log export. If your analytics ritual is a once-a-day glance at visitors and referrers, Matomo's extra depth is not an asset; it is a second database and a gigabyte of RAM spent on reports you will never open.
Umami is cookieless by default — no consent banner needed for basic analytics, nothing to configure. Matomo uses cookies out of the box, so you either reconfigure it to run cookieless or lean on its built-in consent and anonymization tooling. That tooling is genuinely good — it exists because Matomo serves organizations that must track and comply — but it is one more thing to set up correctly, where Umami's posture is simply the default.
The gap here is the widest row in the table. Umami is app plus one database: one volume to back up, a memory footprint that fits the cheapest VPS tier, and effectively nothing to maintain once it is running. Matomo is a classic PHP deployment — the app container plus MariaDB — with twice the RAM floor and the ongoing care a bigger product implies: archive processing, plugin updates, a larger database growing underneath.
Both ship Docker Compose setups and neither is fragile. But "stand it up and forget it exists" describes exactly one of them.
If you want something between these two poles — more reporting than Umami, less machinery than Matomo — the middle ground is Plausible, which we compare against both.
Umami is content on the smallest box a provider sells; Matomo wants 2 GB from day one and is the one to size up as traffic and history grow — MariaDB and Matomo's archive jobs both appreciate headroom. Whichever you pick, back up the database, because that is where every visit you have ever recorded lives. Step-by-step setups for Matomo and Umami are linked below, and any VPS pick here handles either comfortably.
Umami lists a lower minimum: 1024 MB vs 2048 MB for Matomo.
No. Matomo is GPL-3.0-or-later, Umami is MIT.
Pick Umami for the analytics most sites actually need: traffic reports plus funnels, retention and replays, without e-commerce, on a 1 GB box with nothing to babysit. Pick Matomo only when you need Google Analytics-grade depth — goals, e-commerce tracking, raw data ownership — and are willing to run a heavier stack for it.
Umami, decisively. It is a Next.js app with a single PostgreSQL database that runs on 1 GB of RAM at a 2-out-of-5 difficulty. Matomo wants 2 GB and a MariaDB database alongside its PHP app, and rates a 3.
Usually, yes. If you check visitors, referrers, and top pages, Umami's dashboard covers all of it at half the resource cost. Matomo earns its weight when you run a store or a marketing operation that lives in goals, segments, and e-commerce reports.
Umami is cookieless by default, so no banner is required. Matomo uses cookies out of the box — you either configure it to run cookieless or use its built-in consent tooling to stay compliant.
Paid link — we earn a commission if you shop through it.
The full GA replacement vs. the privacy-first dashboard.
Funnels, journeys and replays on both — two containers on 1 GB, or six on ClickHouse and 2 GB.
A single Go binary on a SQLite file vs. the lightest modern analytics dashboard.