Download websites that use Cloudflare Browser Insights

About Cloudflare Browser Insights

Nobody added this script tag by hand. The beacon for Cloudflare Browser Insights is injected by an edge network as the response passes through it, so a detection describes a toggle flipped in a provider dashboard rather than a decision made inside a codebase. No repository holds the code, no reviewer approved it, and the difference is the whole reason the record is interesting.

Cloudflare Inc. ships the feature under a Proprietary model as part of its dashboard, and we file it under Analytics. The beacon collects timing from real visitors rather than from synthetic tests: how long a page took to become useful, how quickly the first byte arrived, how much the layout moved while loading. Dating from 2019, the tool belongs to the generation of monitoring built after lab testing stopped being convincing on its own, and it reports back to the site's own dashboard rather than into a separate analytics account.

As of October 2026 we see the beacon on 2,675,493 sites. United States accounts for 2,230,989 of them and .com carries much of the suffix mix. Concentration on that scale is not a statement about which countries care about performance. Concentration follows the shape of the provider's own customer base, which is weighted toward the same country and the same namespace, so the geography of the beacon is really the geography of the platform underneath it.

For a researcher the useful inference runs sideways rather than downward. A site carrying the beacon is proxied through the provider, sits on a plan where the feature is available, and has an operator who opened a dashboard and deliberately turned something on. Presence therefore separates the accounts that configure their edge from the much larger group that points DNS at it once and never returns, and that separation is hard to obtain any other way. Competing monitoring vendors read the same field as a displacement signal, since a site already collecting real user metrics has admitted the problem matters to it and has not yet paid anybody for a deeper tool. Consultants read it as a qualifier before a performance pitch.

A per-domain check answers whether one site has a toggle switched on, which is rarely worth the query. The downloadable list answers what the toggle correlates with: which platforms, shop systems, and analytics stacks appear beside it, how the population splits across countries and domain suffixes, and how large the addressable segment actually is before anyone builds a campaign around a guess. Aggregation is also the only way to watch the population move, since the interesting fact about an edge feature is usually its adoption curve rather than any individual account.

The limits are straightforward and worth respecting. Detection depends on the beacon appearing in the response our crawler receives, and edge features can be enabled per hostname, disabled on some routes, or suppressed for automated clients, so absence is weak evidence about any single site. The provider has also renamed and reshaped its monitoring products more than once, and older and newer variants do not always look alike inside a page. The population we publish should be read as the sites where a recognizable beacon appeared on a crawled page, which is a floor, and not as every account with the feature quietly enabled somewhere in a dashboard.