The error-tracking docs describe Sentry, the tool we actually use

Three pages of the technical docs described a system that files a Linear ticket for every error, an error endpoint, two Linear settings and console messages confirming the report. None of it existed, so anyone following those pages looked in the wrong tool and trusted a safety net that was not there. The pages also published a Linear intake email address on the public docs site.

The Error tracking page now says what is true:

• Only the web app reports errors, to Sentry. The agent, console, marketing, docs and help sites do not; their errors are in Vercel’s runtime logs.
• Three settings connect it — one for the server, one for the browser, one used at build time.
• What reaches Sentry: server crashes, the error pages, the few places that report a problem on purpose, a sample of traces, and a session replay only around an error, with text masked.
• Where to look: the error page shows a Reference. The same reference is on the Sentry event and the server log, so a user’s screenshot leads straight to both.

The quick-reference page is gone, the Linear address no longer appears anywhere in the docs, and the technology stack and platform overview now name Sentry as the error tracker.