Download websites that use PIXIjs

About PIXIjs

When PIXIjs loads on a page, somebody decided the page needed animation, sprites or an interactive scene that ordinary document markup cannot produce, and a decision of that kind usually arrives with a budget and a specialist to spend it. Analytics, tag managers and font loaders record how a site was assembled. A rendering engine records what the site was meant to do.

PIXIjs is a JavaScript graphics written in JavaScript, published by GoodBoyDigital under the MIT license and in circulation since 2013. Rendering runs through WebGPU where the browser supports it and through WebGL otherwise, so the graphics work is pushed onto the GPU rather than onto the layout engine. Typical hosts are campaign microsites, browser games, product configurators, interactive editorial pieces and playable advertising. Presence of the engine is a rough filter for sites where the front end is the product rather than a wrapper around text.

Segmenting on that filter is worth more than it sounds. Creative and interactive agencies use the population to see who builds this kind of work and for which brands. Advertising technology firms use it to find publishers and studios producing playable formats. Tooling and infrastructure vendors, from asset pipelines to performance monitoring, use it because a site pushing GPU rendering has problems that plain pages never develop. Business development and hiring teams use it because the engine implies skills that are not evenly distributed across the market.

The honest caveat is that detection favours simpler builds. A page loading the engine as its own script, from a content delivery network or a recognisable filename, is straightforward to identify. A project that bundles and minifies everything into a single application file can leave no reliable trace, and the teams most likely to bundle are the teams doing the most ambitious work. As of October 2026 we record 1,798 sites, and the true figure is higher, with the missing part concentrated at the sophisticated end. Our records also carry no version breakdown here, so the file cannot separate installations on recent releases from those on much older ones, which is a real gap given how much the interface changed over the life of the engine.

A per domain lookup tells you that one site uses the engine, which is rarely the question anyone actually has. The useful question is which sites use it, how they cluster by country, top level domain and industry, what else runs alongside, and whether the population is growing between snapshots. Answering that needs the whole set in one file.