Privacy is the inspection step, not a footnote.

Capture starts from an explicit report action, lasts for the report window, and is reviewed before anything is submitted.

Incident recorder, not a visitor recorder.

ClipInbox captures the browser context around a report the user chose to send, not a permanent record of every visitor’s behavior.

Capture boundary
CapturedNot captured by default
Console logs and warningsInput values
Failed request URLs and status codesRequest or response bodies
URL path, title, and referrerCookies, auth headers, and browser storage
Browser, OS, viewport, and visible screenshotFull DOM snapshots or background behavior outside the report session

Why ClipInbox does not record sessions

Session replay can be useful when a team needs a continuous timeline of every visitor action. ClipInbox is built for a narrower job: a person decides to report a bug and sends the browser context needed to reproduce that bug.

Replay tradeoff
QuestionClipInbox answer
Who starts capture?The reporter opens the widget. Capture is not always on.
What is the debugging artifact?A report packet with note, screenshot or clip, console signals, failed request metadata, browser context, and selected recent steps.
What is intentionally missing?No full replay timeline, no full DOM snapshots, no input values, no cookies, no auth headers, no request bodies, no response bodies, and no storage values by default.
When is replay a better fit?When the team needs passive product analytics, broad funnel observation, or a continuous record of behavior before the user decided to report anything.

User-visible by default

The report flow is meant to make the boundary obvious to the person sending it.

  • User-triggeredThe report starts when a user opens the widget. ClipInbox is not a passive surveillance layer.
  • Short-livedCapture is scoped to the report session instead of an endless replay timeline.
  • Review-before-submitUsers can inspect the note, screenshot, and capture summary before sending.
  • Developer-usefulTeams still receive enough browser context to diagnose the issue and reduce follow-up questions.

Give support a better intake path and engineering reports they can use.

Start with a script tag. Let users review the report. Send your team a compact browser-context packet instead of another vague ticket.