Nuxt: 115 sites at the light end of the stack set
Nuxt is a framework for building websites with Vue. It renders pages on a server or at build time, sends them as finished HTML, then hands control to Vue in the browser. Developers meet the name in a job advert or a build log; anyone else meets it because a page's source contains a Nuxt payload. It was detected on 115 of the 5,294 sites measured, 2.2% of them, and those pages sit at the light end of almost every measure in this index's stack section.
The light end of the table
The median Nuxt site contacted 4 companies over 25 requests, with 1 of them classified as tracking. Among the thirteen frameworks and platforms measured here, nothing is lighter on either measure — and Angular reports the same 4 and the same 25, so the bottom of both columns is a shared position rather than Nuxt's alone.
The consent-window figures point the same way. The median Nuxt page had 3 companies contacted before the window closed — nothing among the thirteen is lower, though Angular and Vue report the same 3 — and 85.2% of Nuxt sites recorded any early contact at all, which is the lowest share of the thirteen and is Nuxt's alone.
That is about who picks it, not what it does
Nuxt renders pages. It has no view on which marketing scripts a team adds afterwards, and nothing in it prevents a single line of GTM from adding thirty companies to a page.
What the numbers describe is the population. The sample here is registrars and hosting companies (reg.ru, selectel.ru, name.com, it.com), developer tooling (gitlab.com, dynatrace.com, adjust.com), and product companies with restrained marketing sites (nintendo.com, asus.com, wps.com). Those organisations tend to run few third parties on their own sites: they are selling to people who read network logs for a living. A framework chosen mainly by engineering-led companies inherits the third-party habits of engineering-led companies, and the exceptions in the sample prove it works the other way too — magnit.ru, a supermarket chain, contacted 14 companies with 12 of them inside the early window, on the same framework as gitlab.com's 3.
Early contact is still the normal case
Thirteenth of thirteen still means 98 of the 115 sites had a company contacted inside the early window. Ranking last in a set is not the same as standing outside it.
The median time to first contact was 769 ms, second-highest of the thirteen groups, and the fastest single first contact was 120 ms. selectel.ru shows the crowded version: 10 companies, all 10 inside the window, first contact at 244 ms. palmplaystore.com shows the other one: 3 companies, none inside the window, first contact at 2,213 ms.
Two clusters in one sample
Split the twelve sampled Nuxt sites by where they are run from and the group falls apart cleanly. The Russian-language half — reg.ru with 1 company, selectel.ru with 10, magnit.ru with 14 — showed no detectable consent tool at all, and on selectel.ru and magnit.ru most of the companies were contacted inside the early window. The internationally-facing half — gitlab.com, name.com, dynatrace.com, nintendo.com, adjust.com — each showed one, and their totals run from 3 to 5 companies.
The median for the whole group, 4 companies, sits between two quite different page designs rather than describing a typical Nuxt site. That is worth remembering whenever a framework's numbers look tidy: 115 sites is enough to produce a stable median and not nearly enough to make it mean one thing.
An unusually mixed set of banners
38 of the 115 Nuxt sites carried a consent tool the crawl could detect. Five different tools were seen alongside them: an IAB TCF banner on 18 sites, OneTrust on 12, Cookiebot on 5, CookieYes on 2 and Termly on 1.
Five suppliers across 38 sites is a flat distribution — no single tool takes even half. In the sample, OneTrust appears on gitlab.com, name.com, dynatrace.com and nintendo.com, Cookiebot on adjust.com and a TCF banner on palmplaystore.com. The rest of the sampled sites showed nothing detectable at all.
| Site | Companies | Before consent | First contact |
|---|---|---|---|
| reg.ru | 1 | 1 | 978 ms |
| selectel.ru | 10 | 10 | 244 ms |
| gitlab.com | 3 | 2 | 463 ms |
| name.com | 4 | 4 | 420 ms |
| dynatrace.com | 4 | 4 | 413 ms |
| palmplaystore.com | 3 | 0 | 2213 ms |
| nintendo.com | 3 | 3 | 1077 ms |
| magnit.ru | 14 | 12 | 855 ms |
| asus.com | 5 | 3 | 1183 ms |
| wps.com | 4 | 4 | 941 ms |
| adjust.com | 5 | 5 | 531 ms |
| it.com | 4 | 2 | 1079 ms |
| Tool | Sites |
|---|---|
| IAB TCF consent tool | 18 |
| OneTrust | 12 |
| Cookiebot | 5 |
| CookieYes | 2 |
| Termly | 1 |
Questions
Does building with Nuxt reduce tracking?
No. Nuxt decides how pages are rendered, not which third parties a team adds. The low figures here describe the companies that choose Nuxt, and the same framework will happily serve a page contacting thirty companies.
Is Nuxt the lightest framework in this index?
It is joint-lightest rather than uniquely so. No group of the thirteen measured sits below it on median companies or on median requests, and Angular reports exactly the same 4 companies and 25 requests.
What does the median Nuxt page contact?
4 companies over 25 requests, with 1 company classified as tracking and 3 companies contacted before the consent window closed.
How many Nuxt sites showed a consent banner?
38 of 115. An IAB TCF banner was seen on 18, OneTrust on 12, Cookiebot on 5, CookieYes on 2 and Termly on 1 — some sites carry more than one, so those counts are not slices of a single pie.