Download websites that use Stimulus

About Stimulus

Most front-end frameworks tell you what a site does. Stimulus tells you something about the people who built it. The framework attaches behavior to server-rendered markup through data attributes instead of taking the page over, which is a deliberate refusal of the single-page application model, and teams do not arrive at that position by accident. Detection here is closer to reading a philosophy than to reading a dependency list.

What follows from the philosophy is a fairly predictable stack. A site using it renders HTML on the server, very likely ships a monolith rather than a separate front end and back end, and often pairs the framework with the navigation library from the same family. The team is usually small enough that one person can hold the whole application in their head. None of that is directly observable from outside, which is exactly why the marker is useful: the back end of a website is mostly opaque to a crawler, and a front-end choice this opinionated is one of the few reliable ways to infer what sits behind it.

As of September 2026 we detect Stimulus on 1,467,294 sites, and United States accounts for 1,372,287 of them. Concentration at that level follows a community rather than a technology. The framework came out of 37signals, a product company rather than a platform vendor, and it spread through a professional network of developers who read the same writing and work in the same web framework tradition. Adoption tracks where that network is dense, not where the tool would be technically appropriate, and the suffix mix leans toward .com because the community skews toward commercial products rather than institutional sites.

Released in 2018 under MIT terms, the framework also puts a floor under the age of the code that carries it. A page using it was written recently enough that server-rendered HTML had become a deliberate choice again rather than a default nobody questioned.

The limits are worth stating. Identification relies on controller attributes or the runtime being visible in delivered markup, so a site that compiles everything into a renamed bundle can use the framework without appearing here. The library is small and frequently loaded next to other things, and a site may use it on one screen and nowhere else, which the record cannot express. The total is a floor drawn from what pages expose.

For a recruiter, a developer tool vendor or an agency looking for peer companies, one-domain lookups are the wrong instrument. The question is never whether a particular site uses the framework, it is which companies staff teams that work this way, in which country, at which scale, and which agencies produce this shape of build repeatedly across a client portfolio. A downloadable list of every domain we can identify, with country and suffix attached, answers that in one pass and lets you compare snapshots to see whether the approach is gaining ground. A lookup service, queried one domain at a time, never assembles into a population.