thrumeThe bug is on screen.
The clues come with it.
“Can you send the logs?” “What happened just before that?” Record the reproduction with your selected debug sources, so engineering can watch the failure and inspect the evidence beside it.
Start with what actually happened.
See the setup, the clicks and the result in one reproduction. The person investigating gets the path into the problem.
Put the failed request beside the failed action.
Selected technical sources add another view of the same moment. A visible error and a captured request give the investigation a sharper starting point.
Give the next person somewhere useful to look.
Carry the reproduction, expected result and captured evidence into the handoff. Your team—or an authorized assistant—can decide what to check next.
Bring your
people together.
“The invitation never arrives.”
00:24 POST /api/invites → 500
Inspect the server’s invitation error path.
The moment.
The evidence beside it.
Explore a failed invitation. Each step brings the visible behavior and its available context into the same view.
Your people.
Your next chapter.
Give the team a shared place to work.
Team invitation form is open.
The report includes the steps before the failure, so the next person knows where to start.
Your people.
Your next chapter.
Invitation could not be sent.
POST /api/invites → HTTP 500
The failed request and the visible error happen at the same moment. That is a useful correlation to investigate.
Your people.
Your next chapter.
Invitation could not be sent.
Expected: invitation sent. Actual: no confirmation.
Share the reproduction, captured evidence and expected result. A teammate can inspect the same point—or retrieve it through MCP.
Synthetic walkthrough. Sources must be connected before capture. A matching timestamp helps focus investigation; it does not establish a root cause.
Show the step a written report misses.
A click, a pause, an unexpected result. Record the path to the issue, explain what should have happened and mark the moment it goes wrong. The next person can follow the sequence you actually took.
Bring the browser and the local work together.
A bug can cross the browser, a Mac app and a running command. Connect and select the relevant sources before you record in Debug mode, then keep their timestamped evidence with the take.
- Browser console and supported request metadata through the extension.
- Available logs from the Mac applications you select.
- Selected log files and wrapped command output through the local bridge.
- Screenshots and marked moments that point to the visible failure.
Turn a timestamp into a place to investigate.
Watch the failed action, inspect the nearby events and share the same moment with engineering. Through MCP, an authorized coding assistant can retrieve the captured evidence and help work out what to check next.
Choose the evidence before the take.
Select the sources for this reproduction and review the result before sharing. Source availability varies by app; browser evidence covers supported request metadata. Debug capture starts with your selected, connected sources—it cannot add logs to an earlier Normal recording.
Bring the bug.
And the evidence.
Native Mac recording.
Shared understanding, everywhere.