Download websites that use Catberry.js

About Catberry.js

Rendering the same JavaScript on a server and in a browser was an open problem in the middle of the last decade, and a number of Node projects attacked it before the mainstream frameworks absorbed the idea and turned it into a configuration flag.

Catberry.js was one of those attempts. The framework organised an application into independent components that own their state and their rendering, and it treated server side rendering as the default rather than an add on, so a page arrived as markup and then took over in the browser. Work on it dates to 2014, when the isomorphic label still described a specific engineering problem rather than a marketing category. Released under the MIT license and written in JavaScript, the project lived on npm and in the open, which is how a framework with no company behind it gathers users at all.

As of October 2026 the crawl matches Catberry.js on 1 site. A count that small is the interesting part rather than a disappointing one, because it records what happened to a whole generation of frameworks that solved a real problem and were then overtaken by projects with more people behind them. Sites still carrying the signature are rarely recent builds.

For a researcher the practical meaning of a match is maintenance risk. A page still served by an unmaintained framework implies a codebase nobody has rebuilt in a long time, dependencies pinned to old Node releases, and a team that either cannot or will not touch the front end. Agencies pitching replatforming work, security researchers hunting stale dependency trees and anyone studying framework churn all want the same thing from a page like this, which is the domains rather than the total.

Detection here is inferential, and that deserves saying out loud. The signature we match is a client side artefact, so a build pipeline that renames or inlines it hides the framework completely, and an application that renders only on the server can use it without leaving anything a crawler could see. The figure is therefore a lower bound, and probably a soft one.

A per domain check answers whether one site you already suspect uses the framework. A downloadable list answers the question that actually matters for a survey of ageing front ends, which is where the remaining installations are. No amount of checking domains one at a time will reconstruct that, because the domains running an obscure framework are exactly the ones you would never think to try.