Refresh cadence: roughly alongside each full-census run (historically monthly). The
category/vendor tables (category-prevalence.csv, cdn-vendors.csv, cms-platforms.csv,
ecommerce-platforms.csv, web-servers.csv, ssl-issuers.csv, dns-providers.csv,
tld-distribution.csv, analytics-tracking-load.csv, captcha-crosscheck.csv) are regenerated
by a script hitting the public REST API. rank-band-stack-inversion.csv, cn-visibility-gap.csv,
and sample-domains-10k.csv require direct access to the underlying index (the public API has no
rank-range filter) and are refreshed manually — expect these three specifically to lag the
others by a wider margin between refreshes. This is not an "always current" dataset; for current
data, query the live API.
Source: run_id=top1m-2026-08-25-full-v2, detector version 2026.10.47, 1,000,000 domains
(847,491 reachable, 84.7%). First publication of this dataset. Companion to
We Fingerprinted the Top 1,000,000 Websites.