Download websites that use Public CMS

About Public CMS

The language a content management system is written in usually matters less than the way it is deployed, but Java is the exception. Running a Java CMS means running an application server, a build and deployment process, and an operations team that understands both, which is not a commitment a small site makes casually. Public CMS sits squarely in that world, and the population of sites running it inherits the same character.

The platform is published under Apache 2.0, and our records reach back to 2013, so the code is free to take and free to modify and there is no vendor account sitting behind an installation. Free licensing combined with a heavyweight runtime produces a particular kind of user: organizations with in house developers who want to own the stack, keep the deployment inside their own infrastructure and audit what runs on it. Institutional and enterprise sites fit that description more comfortably than small commercial publishers do.

There is a second reason the language is worth dwelling on. A Java deployment normally sits behind an application server and a reverse proxy, and both leave traces of their own, so a site identified this way tends to arrive with a readable picture of the surrounding stack rather than as one isolated fact. For anyone mapping infrastructure inside larger organizations, the CMS is an entry point into that picture rather than the finding itself.

Version data on this platform is unusually legible. We record 14 distinct versions in use, and the value that recurs is 4.0.202107, a string that carries its own build date inside it. A scheme like that removes the guesswork from the question everyone actually wants answered, which is how long ago a given installation was last updated. Reading a build date straight off a version string, across a whole population of sites at once, is not something a per domain lookup will ever give you.

One caution belongs on the name. Public CMS is a descriptive phrase as much as a product name, and detection rests on fingerprints in the delivered markup and asset paths rather than on the words a site uses about itself. Where templates have been restyled heavily, or a caching layer rewrites those paths, a match can fail even though the software is running underneath. The visible population therefore skews toward installations that stayed near the stock configuration.

As of October 2026 we detect it on 54 sites. Populations of that size reward a different question from the mass market platforms. Nobody needs a dataset to know that generic blogging software is widespread, but a Java CMS with a defined install base is exactly the case where the identity of the sites matters more than the total: which sectors run it, whether the deployments cluster around particular hosting, and how many are serving builds old enough to justify a security or migration conversation. Answering any of that needs the whole set in a file rather than a lookup one domain at a time.