Download websites that use History

About History

Nobody chooses History the way they choose a framework. The library reaches most of the sites it runs on as a dependency of something else, pulled in by a router or a build step and shipped inside a bundle the site owner never inspected. Reading its presence as a preference would be a mistake. Reading it as a fingerprint of how a page was built is exactly right.

What the fingerprint says is that the page is a client rendered application managing its own navigation. History exists to drive the browser session stack from script, so that moving between views updates the address bar and the back button without a full page load. A site carrying it is a site where routing happens in JavaScript rather than on the server, which has consequences for anyone crawling it, measuring it, or trying to make it indexable. As of October 2026 we see History on 447 sites in our Miscellaneous listings.

Version data is where a dependency like this gets interesting. We record 6 distinct versions in circulation, and the one we observe most often is 4.7.2. A spread that wide across a library this small is a sign of bundles built once and never rebuilt, because a transitive dependency only moves when the package above it moves and the site is redeployed. For anyone tracking how quickly real deployments absorb upstream releases, unglamorous libraries are better evidence than headline frameworks, which get upgraded for reasons that have nothing to do with dependency hygiene.

Detection here is inferential and it undercuts itself. We identify the library from artefacts that survive bundling, and a build that mangles names or inlines dependencies aggressively can hide it completely. The code is BSD licensed and distributed through package registries, so there is no vendor telemetry to check our count against. The list is best treated as a floor rather than a census.

A floor is still useful, because it is a list rather than a lookup. Checking one domain tells you what one team shipped. The export tells you which populations of sites share a build lineage, which lets a security team scope who might be affected by an advisory against a version range, lets a tooling vendor find the applications their product is meant for, and lets an analyst compare routing adoption across market segments instead of guessing at it from developer surveys. Dependency questions are population questions, and per domain answers cannot be added up into one.