Download websites that use Facebook Pixel

About Facebook Pixel

Almost everything else we detect tells you how a site was built. Facebook Pixel tells you that somebody is spending money. The snippet exists to report visitor behavior back to an advertising platform so campaigns can be optimized, audiences rebuilt, and conversions attributed, and nobody installs it without an account and a budget standing behind it.

Introduced in 2015 by Meta under a Proprietary arrangement and filed here under Analytics, the snippet fires named events as a visitor moves through a page: a view, a product opened, a cart addition, a completed purchase. Each event travels back to the advertising platform attached to an identifier, which is how a retargeting audience gets assembled and how the return on a campaign is calculated. Nothing about the mechanism is subtle, and nothing about it is accidental.

As of September 2026 we find the snippet on 2,326,510 sites, with United States accounting for 1,095,142 and .com carrying a large share of the volume. Geography here is genuinely confounded, and reading it as a map of advertiser density would be a mistake. Consent frameworks in some regions prevent the snippet from loading until a visitor accepts a banner, so a crawler that does not click never sees it, and the regions with the strictest consent rules end up underrepresented for reasons that have nothing to do with how much advertising happens there.

The version field behaves unlike a normal dependency. We hold 264 distinct values, with 2.9.256 appearing widely, and yet no site owner ever chose any of them, because the tag loads its real payload from the platform and the value reflects whatever build was being served when a page was crawled. The column is a rollout timeline belonging to the vendor rather than a maintenance signal belonging to the site, and treating it as evidence of neglect or diligence would misread it completely.

The audience for the aggregate becomes obvious once the signal is understood. Agencies prospect against it, because a site running the tag has already demonstrated that it buys traffic and does not need convincing that advertising works. Competitive analysts use it to see which categories a rival is advertising into and what other stacks sit beside the tag. Privacy researchers and regulators use it to measure how far a single tracking infrastructure reaches across an entire national web, which is a question only a full population can answer and never a single lookup.

A per-domain check is a devtools exercise. The downloadable list answers the questions that need the whole set at once: how the advertiser population splits by country and by domain suffix, which shop platforms and content systems it runs on, how it overlaps with the other advertising and analytics tags in the same records, and how a segment can be sized before anybody prices a campaign against it.

The undercount is the caveat that matters. Many sites now load the tag through a tag manager, which can hide it from a static scan of the page source, and a growing number report conversions from their own servers instead, leaving nothing in the page at all. Presence is strong evidence that a site advertises. Absence is weak evidence that it does not, and any analysis built on the second reading will be wrong in a direction that flatters whoever is measuring.