What do design studios use to build their own websites? We reviewed an export of automated detections from 400 studio sites in the D1S1 directory. The most useful finding is how tools combine: a website builder can sit alongside custom animation, and a framework can pair with a separate CMS.
Snapshot: March 10, 2026. Reviewed September 4, 2026. These figures describe the original 400-site export. The technology directory shows counts from the latest stored detections, which may differ as studios are added or rescanned.
The review corrects the original article's Webflow–GSAP count and withdraws its Rive ranking, growth claims and claims about how much code studios write. Those conclusions were not supported by the recorded evidence.
How to read the numbers
The sample is a curated directory, not a representative survey of the design industry. It describes studios' own portfolio websites, not the tools they use for client projects or their development capabilities.
Each technology is counted once per studio. A site can use several tools, so categories overlap and percentages must not be added together. All overall percentages use 400 as the denominator; combinations below use the named platform's count.
Detection finds clues in fetched pages, scripts, stylesheets and response headers. It can miss bundled or hidden tools and can produce false positives. The recorded confidence scores describe detector rules; they are not measured probabilities of correctness. No detected platform means unknown, not proof of a custom build.
WordPress leads the recorded detections
WordPress appears on 155 sites. Of those, 151 have WordPress-specific asset paths; four were tagged only because their server reported PHP, which is insufficient to identify WordPress. Even excluding those four, WordPress remains ahead of the other named platforms in this export.
Semplice, a portfolio system built on WordPress, appears on 18 sites, including Stupid, Moniker and Hyperfocus. It should not be counted as a separate, mutually exclusive alternative to WordPress.
GSAP also appears on 32 of the 155 WordPress-tagged sites (21%). That coexistence is useful when researching a portfolio build: a CMS choice does not determine the entire animation stack.
Webflow and custom animation coexist
Webflow appears on 49 sites, all with its distinctive page attributes. 19 of those 49 also have GSAP detections (39%). The original article reported 22; recounting the saved export gives 19.
This is a useful set of references for anyone considering Webflow with additional animation. It does not tell us who wrote the animation code, how complex it is, or whether the result performs well.
Framer is present, but growth needs another measurement
Framer appears on 20 sites (5%), including Bleed, Goods, BAS ID, Shuka and IMCOLLECTIVE. All 20 have Framer-specific HTML attributes in the saved evidence.
That establishes adoption within this sample. A single snapshot cannot establish that Framer is growing fastest, or that studios switched to it from another platform. To measure that, we need repeated scans of the same studios with recorded dates and platform changes.
Other recorded choices include Squarespace on 18 sites, Readymag on five and Cargo on four. These are examples to explore, not a verdict on which tool is best.
Next.js frequently pairs with Sanity and Vercel
Next.js appears on 61 sites, compared with 19 for Nuxt and seven for Gatsby. Across Next.js, Nuxt, Gatsby and Astro, 90 unique studios have at least one detection (22.5% of the sample). Framework detection does not imply that a site was written from scratch.
Within the 61 Next.js sites:
The complete three-tool combination appears on 21 sites, not every Next.js site. These pairings offer a practical starting point for researching a framework, content management and hosting setup. They do not establish an industry standard or explain why each studio chose it.
Why we withdrew the Rive ranking
The original article called Rive the most popular animation tool, ahead of GSAP. That claim does not survive an evidence review.
Of 73 recorded Rive detections, 72 relied only on a loose rive text match. That sequence can appear inside unrelated words or script content. Just one detection had a more specific rive.app signal. This does not mean only one studio used Rive; it means the export cannot establish the true total or support a popularity ranking.
A stronger detector should look for runtime packages, initialization code and animation assets together. Rive's official web runtime documentation illustrates package names such as @rive-app/webgl2, the rive.Rive constructor and .riv assets. Even a specific signal should be checked in context.
What the scan cannot tell us
The original “64% build without writing code” claim came from 258 sites with at least one platform in a CMS/builder grouping. The grouping includes WordPress and overlapping tools. It does not measure whether anyone wrote code.
Similarly, 62 sites without a detected platform or framework should be labelled unknown, not “fully custom.” Missing award badges do not tell us whether a studio enters competitions or has won awards.
The scraper's page-weight field measures the fetched HTML response, not the total size of images, video, scripts and fonts. Its response time is not a browser loading or interaction metric. Neither supports a conclusion about the full site's performance.
How to use this research
Start with a build question: which studios use Webflow alongside animation tooling, or Next.js alongside a separate CMS? Follow the relevant technology directory links, then inspect the studios' websites as references. Browsing studio lists by technology requires Pro; this analysis is public.
The strongest next step for the dataset is better evidence: stricter detection rules, visible scan dates, scan coverage and repeated measurements of the same studios. That would let us distinguish reliable adoption counts from detection noise, and actual platform switches from growth in the directory itself.
