Captured context

web

Exactly what ships with a report, with every cap spelled out. If it isn't listed here, it doesn't leave the page.

Always captured

optiontypedefaultdescription
consolering buffer100 entrieslog / info / warn / error, each truncated to 500 chars. Captured from init onward.
networkring buffer50 requestsfetch and XHR: method, URL, status, duration. Request and response bodies are never read.
errorsring buffer20 entriesUncaught JS errors and unhandled promise rejections, with file:line when available.
navigationbreadcrumbs30 routespushState / replaceState / popstate trail — how the tester got here.
environmentobjectViewport, browser, language, timezone, user agent, connection type.
DOM snapshotgzipped≤ 800 KBrrweb snapshot of the page, inputs masked. Rebuilt in the dashboard, sharp at any zoom. Skipped if over the cap.

Captured on demand

optiontypedefaultdescription
screenshotJPEG≤ 3 MBOnly when the tester uses Capture & annotate. Blur regions are pixelated before upload.
annotationsshapes[]≤ 200Vector annotations: brush (point list), arrow, rect, blur — with coordinates and color.
session recordinggzipped≤ 3 MBOnly during a test session. rrweb event stream across page loads, merged at send time.
click journeylist≤ 200Only during a test session. Timestamped clicked elements with their visible label.
noteThe whole report is capped at 8 MB; individual media that would blow the cap are dropped rather than failing the report. The technical context always gets through.

What is never captured

  • Request or response bodies — ever.
  • Input values — masked in both the DOM snapshot and session recordings; passwords are always masked.
  • Cookies, storage contents, auth headers.
  • Anything under a tester's blur annotation.