Download websites that use Wurfl

About Wurfl

Serving the right page to a phone meant first identifying the phone, and that identification problem is what Wurfl was built to solve. The library reads the headers a browser sends, chiefly the user agent string, and looks them up against a maintained database of device models and their capabilities, returning screen dimensions, input methods and format support that a page can then adapt to. Our record credits the work to ScientiaMobile, dates the project to 2001 and lists Java as its implementation language, with licensing recorded as Freemium.

As of October 2026 we see Wurfl on 2 sites. A count that small deserves an explanation rather than a footnote, because the classic deployment of a device capability database is invisible from outside the server.

Detection depends on something arriving in the browser. When capability lookup runs server side, in a Java filter or a module sitting in front of the application, the visitor receives ordinary adapted HTML and nothing in that page identifies the mechanism that produced it. The sites we can attribute are the ones loading a hosted client side variant, where a script tag or a beacon points at the vendor domain. Our number is best read as a floor under the hosted flavour, not as a census of everyone doing device detection.

The classification we carry, JavaScript libraries, says something about who pays for this. Device capability data feeds mobile optimisation, advertising decisions and reporting on the pages where it runs, and the vendor sells the hosted script to the site operator itself, with a free tier for coarse device facts and paid tiers for the detailed model names. A page carrying the script is usually a page where somebody is measuring the audience by hardware, which is a useful signal if you are mapping who still invests in device level targeting rather than leaving adaptation to CSS.

For research the value of a downloadable list is the shape of the set, not any single hit. Checking one domain tells you whether that domain uses it. A full extract lets you ask which industries still run explicit device detection, whether the sites doing it cluster in publishing or in commerce, and what analytics and advertising code sits alongside on the same pages, questions that need the whole population in one file. The honest framing is that we describe a client side footprint, and that a serious study of device detection would have to combine our list with server behaviour tests that no crawler can run at scale.