Download websites that use TrackJs

About TrackJs

Error monitoring is one of the few things in a page that exists purely for the people who built it. Visitors never see TrackJs. A JavaScript agent loads early, watches for exceptions, records the user actions and network calls that preceded them, and ships the lot to the vendor, and none of that changes anything a visitor experiences. The script is in the page because a development team wanted to know when their front end breaks in the wild.

Detection therefore reads as a maturity marker. A site carrying a Analytics agent has developers who own it in production, a release process that outlives the launch, and somebody whose job includes reading the alerts. Brochure sites do not have one. Applications, portals, checkout flows and public-facing internal tools do. When a technology list is used to find companies that build software rather than companies that merely have a website, front end error monitoring is one of the cleaner filters available.

TrackJS has sold the Proprietary service since 2013, positioned narrowly around browser-side errors and the sequence of events that led to them rather than around infrastructure metrics. Narrow positioning gives the customer base a character: teams that hit a JavaScript problem they could not reproduce and went looking for a specialised tool, often agencies and product teams rather than large organisations with an in-house observability stack. Knowing which vendor a site chose is a stronger qualifier than knowing only that it monitors something.

There is a budget signal underneath as well. Monitoring is a recurring cost with no visible customer benefit, so it survives a budget review only where engineering has enough standing to defend it. Sites that carry a paid agent year after year belong to organisations where software is treated as an ongoing operation rather than as a project that finished.

As of October 2026 we detect TrackJs on 591 sites, and the number deserves a caveat running in both directions. Agents are often loaded through a tag manager, or self-hosted from the customer's own domain to survive ad blockers, in which case the vendor fingerprint is weak or absent. Installations confined to staging environments or to authenticated pages we never reach are invisible to us as well. The count is a visible surface, not an account list.

The list is the point. A competing monitoring vendor uses it to build a displacement target set. A developer tools company uses it as a proxy for engineering headcount. A researcher tracking the observability market uses successive crawls to watch installs move between vendors. All of those are population questions, and none of them can be answered one domain at a time.