Companies Using React Router

React Router is a routing library, which sounds like the least interesting thing in a modern JavaScript stack. It turns out to be the most useful measuring instrument we have, for a reason that has nothing to do with routing: a site cannot run React Router without running React. That single dependency makes React Router a ground truth against which every other React signal can be checked.

The headline first. TechnologyChecker.io detects React Router on 392,715 domains, fourth in JavaScript Frameworks at 5.7% of the category. Its version is written into the page, so we read a release number on 99.99% of detections, which is the highest version-capture rate of anything we publish. Version 6 holds 57.34%, version 7 holds 40.12%, and version 8 has reached 2.53%.

That version 7 figure is worth pausing on, because version 7 is where Remix was folded into React Router. The merge was announced as the consolidation of two overlapping projects, and 40.12% of the installed base has arrived at it. As adoption curves for a major version go, that is fast.

What has not happened is the part everyone assumed would follow. We looked at every domain where we detect Remix and asked which React Router version sits alongside it. 86.3% carry React Router version 6, which is the router bundled inside Remix 2. Three carry version 7. Not three percent; three domains. Deployed Remix is still Remix 2, and the migration path into React Router 7 that the merge was supposed to open has essentially no traffic on it yet.

Remix is not a separate routing population. On the evidence of live domains it is React Router version 6 wearing a different name, and the consolidation exists in the repository rather than on the web.

Now the finding that made this profile worth writing. React Router requires React. So on any domain where we detect React Router, React is present by definition. We detect a React signal on 11.99% of them.

The scale of the gap is easier to see historically. Across 344,714 domains that have at some point shown both React and React Router, only 54,542 still show both today. 258,748 have lost the React signal while keeping React Router. Those sites did not remove React; removing React would have taken React Router with it. Our React detection simply stopped seeing them.

We checked this the only way that settles it, by loading pages in a real browser rather than reading markup. On Sanofi, Uber and Apple Card, React Router was present every time, reporting versions 7.18.3, 6 and 6 respectively. On all three, no React global was exposed at all. Three out of three agreed with the corpus.

The mechanism is not mysterious. React's detection leans on markers that modern React no longer leaves behind, most notably a root-container property that disappeared when React 18 changed how applications mount. A site upgrading React does not become less React; it becomes less visible. Corpus-wide, React has lost its signal on 1.31 million domains, a 30.65% drop rate, against 10.57% for React Router and 6.52% for Angular. The largest technology in the category is also the one decaying fastest, and one of those two facts is about the web while the other is about our rule.

React Router puts a floor under React that React's own detection cannot reach: at least 392,715 live domains are running React right now, and only 12% of them say so.

This has a practical consequence for anyone using technographic data on the React ecosystem, ours included. Do not size the React market from a React signal. Size it from the things React cannot run without, or from the things that cannot run without React. React Router is the best of those instruments because its own detection is nearly perfect: a version string on 99.99% of hits leaves very little room for a false positive.

The rest of the profile is a portrait of contemporary React work, and it is strikingly uniform. Tailwind CSS appears on 69.23% of React Router domains, Lucide on 49.99%, Radix UI on 43.92% and shadcn/ui on 39.76%. Those are not four independent decisions. shadcn/ui is built on Radix primitives, styled with Tailwind, and ships Lucide icons as its default set, so a team that runs one typically runs all four. Roughly two in five React Router applications were scaffolded from the same template.

The companies behind them are young and small. 64.52% employ ten people or fewer and only 5.15% employ more than 200, against 13.97% for Angular. 41.42% were founded in the 2020s and just 1.88% before 1960, where Angular's pre-1960 share is 8.72%. Set the two profiles side by side and you get the clearest picture of the split in front-end work that we can produce from crawl data: Angular maintained by old institutions on old versions, React Router adopted by new companies on new ones.

Geographically the United States takes 28.71% and India 11.68%, the strongest Indian share in this batch, with Brazil at 7.19% and the United Kingdom at 6.63%. Germany, which supplies 10.50% of Angular's base, reaches only 3.98% here. Software Development, IT Services and Technology together account for 21.26% of matched companies, against 17.83% on Angular, and Banking does not reach React Router's top ten at all.

The large enterprises that do appear are worth naming because they contradict the small-company averages: IBM, Apple, Ford, Wells Fargo, Uber, UBS, Roche, Medtronic, Morgan Stanley and DHL all serve React Router from production subdomains. As with Angular, the detections land on the transactional edges of the business, on appointments.wellsfargo.com, accountmanager.ford.com and ebanking-de.ubs.com, rather than on the corporate homepage.

We publish no switching figures on this page. React Router's category neighbours are either its own dependency or members of other ecosystems, and every candidate corridor failed once it was measured against how often the other technology's signal disappears everywhere else. A migration claim has to beat that base rate to mean anything, and none of these did.