Download websites that use Docusaurus

About Docusaurus

A documentation site is a different kind of asset from a homepage, and being able to find one is a way of finding the product behind it.

Docusaurus is a Documentation maintained by Meta and released under the MIT license, built on React and pointed squarely at project documentation rather than at general publishing. First shipped in 2017, it became the default answer for teams that needed versioned docs, a search box and a translation workflow without hiring anyone to build them. Because it is a static generator written in JavaScript, the output is plain files that can sit on any host or edge network, which is also why documentation built with it tends to outlast the team that set it up.

A Docusaurus install is rarely a company's primary website. Finding one usually means finding the documentation for a library, an API or a developer product, quite often on a subdomain or a path rather than at the root. For anyone selling to developer facing companies, the docs surface is the more informative one, because a company that maintains public documentation has a product with external users, a writer or engineer assigned to it and a release process worth documenting. Marketing pages say a company is technical. A documentation site shows it.

Version data is unusually legible for this generator. We currently distinguish 93 versions of Docusaurus across the installed base, and 3.9.2 is the build that recurs most often in the records. A generator that ships breaking major releases leaves a long tail of sites pinned to older branches, and the spread between them is a fair measure of how actively each documentation site is maintained. A project sitting on a current build has someone still tending it. A project several majors behind is often one whose maintainers moved on, which is a useful signal for anyone evaluating a dependency, a partner or an acquisition target.

As of September 2026 we detect Docusaurus on 10,506 sites. Detection leans on the generator meta tag and the asset naming a default build emits, both of which a heavily customized theme can strip away, so the count should be read as a floor. Site structure matters too, since documentation is often served from a subdomain that has to be reached on its own rather than through the main domain.

What the export gives you is a population instead of an anecdote. Developer tool marketers use it to assemble a list of companies that publish documentation at all. Documentation platforms and technical writing vendors use it to find migration candidates, especially the stale ones. Researchers use it to watch how quickly a major version propagates through sites that actually exist, rather than through download counters that anyone can inflate. A per domain lookup tells you what one docs site runs, and by construction it cannot tell you which docs sites are out there.