Download websites that use Lightning

About Lightning

Names in a technology dataset are not always unique, and this entry is a case where one label carries more than one product. The vendor recorded against it is Vektor,Inc., whose component framework for its own application platform is the widely recognised bearer of the name. The distribution we actually observe points somewhere else, and a researcher who assumes the enterprise framework will misread every row.

Of the sites where we detect Lightning as of September 2026, 65,577 out of 70,670 resolve to Japan. A component framework used inside a global business application would not concentrate in a single national market like that. A WordPress theme published by a Japanese studio and used mainly by Japanese businesses would, and the release series we read off these installs is consistent with a theme on a fast release cadence rather than with an enterprise platform's internal component library.

The version evidence is unusually rich here. We record 652 distinct releases in the wild, and the one appearing on more sites than any other is 15.37.1. A spread that wide means installs are pinned all across the release history, which is what happens when software updates often and the people running it update rarely. Anyone planning outreach, compatibility work or a security notice against this population needs the release each host is actually on, not the release currently shipping.

Presence on a domain therefore implies a fairly specific kind of site. Somebody built a business site on WordPress, chose a theme with an opinionated block layout whose interactive parts are written in JavaScript, and runs it on ordinary hosting in one national market. Agencies selling redesigns, plugin and form vendors, hosting companies and localisation services all sell into exactly that profile, and the domestic concentration means the list doubles as a market map for a country that is otherwise hard to prospect from the outside. The license recorded is GPL v2, which fits the enterprise framework rather than a theme distributed through the public WordPress directory. Our own counting is done from the pages themselves.

Detection is inferential. We match asset paths, generator markers and class-name patterns in delivered markup, and a heavily customised install, a full-page cache that rewrites asset URLs, or a build that inlines its styles can hide any of that. Naming ambiguity cuts the other way as well. Where a fingerprint could plausibly belong to either product, the label alone will not separate them, and the safe reading of any single row is that the site carries something published under this name. Country, suffix and version together disambiguate a row far better than the name does.

A per-domain lookup returns the label and stops, which is precisely the wrong tool for an ambiguous label. The downloadable list returns the population with the columns that make the label interpretable: .com and the other suffixes, the country resolution, the version string per host, and every other technology found on the same page. Holding all of it at once is what lets you split the population into the group that matches one product and the group that matches the other, rather than taking a single ambiguous row at face value.