Download websites that use Bulma

About Bulma

A CSS framework that ships no JavaScript is a statement about how a team wants to work. Bulma gives you a grid, components and typography as class names and nothing else, so there is no widget lifecycle to learn, no plugin API to fight and no runtime dependency to reconcile with whatever framework renders the page. Teams pick it when they want layout solved and everything interactive left to them.

As of October 2026 we detect it on 9,216 sites, filed under UI frameworks even though the code itself is CSS. 4,492 of those sites are in United States and .com carries the dominant share of the domains, a distribution that follows the developer-tool audience rather than any particular industry. Documentation sites, dashboards, internal tools, small product pages and side projects are where the framework tends to surface.

No company stands behind it. Released in 2016 under MIT and organised as modular Sass files so a project can import only the parts it uses, the framework is maintained as a community open-source effort rather than on a vendor roadmap. For anyone assessing risk in a stack, that governance difference usually matters more than a feature comparison against a commercially backed alternative.

Version data says something specific about the population. 55 distinct versions appear in the dataset and 0.9.4 is the one we see most often, which is a concentration on a mature release line rather than a scatter across recent ones. Frameworks consumed through a copied stylesheet link behave that way. Somebody pastes the tag once, the page keeps working forever, and there is no dependency manager nagging anybody to move. Populations that install through a build pipeline look different, because upgrades ride along with routine maintenance.

The same mechanism shapes what we can detect, and the bias runs in one direction. Recognition depends on seeing the framework's class names and stylesheet in the delivered page, so a project that runs a purge step, renames classes or compiles the Sass into a single custom stylesheet effectively disappears. The sites we count therefore skew toward the plain, unbuilt end of usage, and the real footprint is larger than the number. Anyone modelling adoption should treat the figure as a floor and the composition as tilted toward simpler deployments.

Read with that caveat, the list answers a question that per-domain lookups cannot. Checking one site tells you what one site loads. A full extract tells you how big the population is, where it clusters by country and suffix, which release line it has settled on, and how much of it moves between releases. Component library vendors, template sellers, agencies choosing a house stack, and researchers tracking how quickly the CSS layer of the web turns over all need the population rather than the sample, and the sample is all a lookup can ever return.