Download websites that use Amaya

About Amaya

Not everything a crawler finds is in use. Some of what it finds is a signature left behind by a tool that ran once, years ago, and never came back. Amaya belongs to that second category, and it makes a particularly clear example, because the product was never a web server or a framework. The program was an editor and browser built at W3C to test the standards the organisation was writing.

A page authored in such a program carries the fact in its markup, in the generator metadata that authoring tools write into the head of a document. Our detection reads that signature. So a domain in this list is not running software. The domain is serving a document that was written with a particular editor and has survived, unedited in that respect, ever since.

The dates make the point. Amaya traces to 1996 and was written in C, in the era when the web's own institutions still shipped client software, and the tool is no longer part of anyone's working toolchain. Anything we find today is therefore a page that has outlived its authoring program, the habits of its authors and in many cases its original purpose.

The signature has a known weakness. Generator metadata is copied whenever a document is copied, so a template written once and reused across a site will report the editor on every page that inherited the markup, and a page can carry the tag while its visible content was rewritten by hand years later. Our count records where the marker is present, not how much of the page the tool actually produced. Presence proves an origin, not authorship of everything on the page.

We do hold release detail, which is unusual for a signature this old. A version spread that narrow is consistent with what the signature represents, which is a small and closed population of documents frozen at whatever build their author happened to be running.

The audience for a list like this is not commercial in the usual sense. Nobody sells a plugin to a page written with a retired standards body editor. The value lies in what such pages indicate about the domains that host them, since a document nobody has touched in a decade usually sits on infrastructure nobody has audited in a decade either. Researchers studying web decay, institutional link rot and the survival of early academic sites use exactly this kind of marker to find a sample. Security teams read the same signal in a harsher way, as an indicator of unattended hosting.

As of October 2026 the signature appears on 147 sites, a population small enough to read through but too dispersed to find by guessing. No search engine surfaces pages by authoring tool, and no per domain lookup helps, because the researcher does not know which domains to check. Starting from the full set of matches is the practical way in, and a downloadable file is exactly that: the whole matched population, sortable by country code and top level domain, ready to be narrowed to the corner of the web the question is actually about.