Next.js
Next.js is a React framework. It renders pages on a server or builds them ahead of time, then hands the browser a mostly finished page. Sites do not fetch it from a third party at runtime the way they fetch an analytics script; it ships as part of the site's own code. That is why a detector reporting Next.js tells you something about who built a site and nothing about who else that site talks to. It was detected on 756 of the 5,294 sites this snapshot covers, 14.3% of them, which makes it third of the 13 technologies measured here.
Third by adoption, eighth by weight
Next.js ranks third of 13 technologies in this index by number of sites. By what those sites load it ranks nowhere in particular. Its median site contacted 6 companies (8th of 13), carried 1 tracker (8th of 13), issued 36 requests (7th of 13), and made its first third-party contact at 673 ms (8th of 13). A technology that sits third by adoption and eighth on every load measure is showing you that the two are unrelated. The build tool does not decide what gets added to a page afterwards.
The spread inside the 756 is the interesting part
The example sites recorded for Next.js run from one company contacted to twenty-two. github.io and snapchat.com each contacted a single company. openai.com contacted 3. vimeo.com, nginx.com and f5.com contacted 4. unity3d.com contacted 7. applovin.com contacted 18, with 16 of them inside the consent window. kaspersky.com contacted 22, with 3 inside it. That is a twenty-two-fold range on one framework, and it is the reason a per-site measurement answers a question that a per-framework median cannot.
694 of 756 sites, and one instructive exception
On 694 sites, 91.8% of them, at least one company was contacted before a visitor could have answered a consent banner. Among the 13 technologies here that share is 7th, mid-table again. The median site contacted 4 companies inside that window, and once the window had closed the median site added none: the median count of companies arriving afterwards is 0. godaddy.com is the counter-case in the example set: 3 companies contacted, not one of them inside the window, and nothing leaving the browser until 2,025 ms.
Fewer than half carry a banner the crawl recognises
373 of the 756 sites carried a consent tool the crawl could identify. Alongside Next.js it saw OneTrust on 190 sites and an IAB TCF consent tool on 120, then Cookiebot on 21, Osano on 15, CookieYes on 11 and Usercentrics on 9. Those counts are overlapping observations, not slices of the 373, because a site can carry more than one banner. In the example set, unity3d.com, intuit.com and applovin.com registered OneTrust and one.one registered Cookiebot, while vimeo.com, openai.com, kaspersky.com, snapchat.com and godaddy.com registered none the crawl could name.
What a framework detection is good for
Knowing a site runs Next.js narrows down its build pipeline and its hosting options. It does not narrow down its third-party exposure, because the numbers above cover sites that contact one company and sites that contact twenty-two. If the question is what a specific page loads and when, the answer has to be measured on that page. MurmStack does that in the browser and reports the same fields used here.
| Site | Companies | Before consent | First contact |
|---|---|---|---|
| vimeo.com | 4 | 4 | 320 ms |
| nginx.com | 4 | 2 | 1348 ms |
| github.io | 1 | 1 | 1253 ms |
| snapchat.com | 1 | 1 | 1110 ms |
| unity3d.com | 7 | 6 | 992 ms |
| openai.com | 3 | 3 | 532 ms |
| intuit.com | 9 | 9 | 795 ms |
| kaspersky.com | 22 | 3 | 1460 ms |
| godaddy.com | 3 | 0 | 2025 ms |
| one.one | 5 | 2 | 1298 ms |
| applovin.com | 18 | 16 | 661 ms |
| f5.com | 4 | 3 | 895 ms |
| Tool | Sites |
|---|---|
| OneTrust | 190 |
| IAB TCF consent tool | 120 |
| Cookiebot | 21 |
| Osano | 15 |
| CookieYes | 11 |
| Usercentrics | 9 |
The six most common are listed. The remaining 7 of 373 sites with a banner carry something outside that six. One tool is recorded per site, so these counts never overlap.
Questions
How many sites in this index run Next.js?
756 of the 5,294 sites this snapshot covers, or 14.3%, which is third of the 13 technologies measured here.
Does Next.js itself contact third parties?
It ships as part of a site's own code rather than loading from an outside host at runtime, so detecting it describes how the site was built. What the site contacts is a separate matter: the median Next.js site here contacted 6 companies and carried 1 tracker, and individual sites in the example set ranged from 1 to 22.
How many Next.js sites had a consent banner?
373 of the 756 carried a consent tool the crawl could identify, so a little under half. OneTrust was the one seen alongside most often, on 190 sites.
How quickly does the first third party arrive on a Next.js site?
The median is 673 ms, which is 8th of the 13 technologies here. The earliest first contact recorded on any Next.js site in this snapshot was 84 ms.
Is a lower median better than a higher one?
Not on its own. A low median company count can mean a lean page or a page that simply sells nothing. These figures describe what was contacted and when, not whether a site's choices were appropriate.