Download websites that use ButterCMS

About ButterCMS

A headless content system is designed to leave no trace, which makes it an awkward thing to count. Content is fetched over an API on the server or at build time, assembled into ordinary markup, and delivered to the browser with nothing that identifies where it came from. ButterCMS works that way, so any honest discussion of its footprint has to begin with what the numbers cannot see.

As of September 2026 we detect ButterCMS on 171 sites. Anyone treating that as the size of the customer base will be wrong by an unknown and probably large factor, because the detectable sites are the ones where the integration happens to be visible, through a JavaScript client running in the browser, a preview or draft mechanism, an asset served from vendor infrastructure or a template naming its source. Sites rendering entirely on the server, which is the intended architecture, are invisible to any crawler including ours. The figure is a lower bound with a wide gap under it, and saying so is more useful than presenting a confident number.

What survives that caveat is still useful, because the visible subset is not random. A site detectable as a headless CMS user is one where developers wired the API into the front end directly, which tends to indicate a JavaScript application rather than a traditional template stack, and a team comfortable with content arriving as data. The list therefore skews toward exactly the audience that sells to developers, and away from marketing sites built by agencies on server rendered frameworks.

The commercial reading also differs from that of a monolithic CMS. Detecting a traditional publishing tool tells you a site owner picked software to write in. Detecting ButterCMS tells you a company already had an application and needed to bolt editing onto it, which is a different budget, a different buyer and usually a different scale of organisation. ButterCMS sells the service under a Proprietary model, so there is no self hosted population to confuse with paying accounts, and the product has been on the market since 2014, long enough for the installed base to hold both early adopters and recent arrivals.

Researchers who want this list usually want it for one of three reasons. Competing content platforms want the visible part of a rival's base as a sales target and as a rough read on category adoption. Agencies and consultancies want companies that have committed to an API driven architecture, because those companies buy integration work. Anyone studying the shift away from monolithic publishing wants a repeatable measurement, and a repeatable undercount measured the same way each snapshot still shows direction even when it cannot show magnitude. A single domain lookup does none of that, because the question is about a population and not about one site.