Brave on Google Play — install bucket and review count (Aug 2026)
The Play Store listing for com.brave.browser, read 2026-08-22 (via Firecrawl; the page is
JS-rendered and a plain fetch returns the shell).
T2 — a platform's own count of a third party's app. Google publishes the install bucket and the
review count; Google does not ship Brave. The vendor copy on the same page is T3 and is kept
separate below. volatile: both figures move, and the bucket is coarse enough to hide years of
movement.
The two numbers, and they are not the same kind of number
| Downloads (Google-published bucket) | 100M+ |
| Reviews (Google-published count) | 3.71M |
| "Loved by over 100 million users" | Brave's own listing copy — T3, a vendor claim |
Name the coincidence, so nobody launders one figure into the other. Google's bucket says at least 100 million installs, ever, on Android. Brave's sentence says users. They are different quantities about different populations, they appear on the same screen, and only the first is measured by somebody other than Brave.
Why this is on the edge at all
brave ships a Chrome user-agent by design, so every user-agent panel — statcounter-browser-share, dap-browser-share — counts it as Chrome. homebrew-cask-analytics got around that by counting a package name (93,290 macOS installs in the year to 2026-08-21). This is the second such channel and the first outside macOS: Play counts the package, so the user-agent trick is irrelevant here too.
What it is worth, honestly
It is a floor, an order of magnitude wide, and cumulative. "100M+" is one of Play's coarse buckets — the next one up is 500M+ — so the true figure sits somewhere in a 5× range. It counts installs since launch, not devices still running it, not people, and not anyone who uninstalled. Nothing here is comparable with dap-browser-share's 39,976,619 concurrent users on US government sites or with a monthly-active figure.
The review count is the more informative half. 3.71M reviews is a directly counted number rather than a bucket, it is produced by Google, and it is the kind of quantity that scales with an active base. It is not a user count either, and this wiki holds no ratio anywhere that would convert it into one — which is the honest limit.
Three channels, three orderings. Homebrew put vivaldi ahead of opera 1.9:1 where DAP puts Opera ahead about 1,160:1; Play adds a third instrument that reaches a different population again (Android consumers, worldwide) and cannot be lined up with either. The corpus's standing conclusion holds and gets stronger: a panel is a channel, not a population, and Brave in particular is only ever measured by whoever happens to be counting packages.
Related
brave · homebrew-cask-analytics · dap-browser-share · statcounter-browser-share · vivaldi · opera