Open a transcript for review
When a transcript finishes processing it lands in the Needs review group on the Transcripts page. Click Review on the row to open the review sheet. The review sheet has two panels side by side:- Speakers — the speaker rename controls.
- Transcript — the full diarized text.
The transcript review sheet: hover a segment to see the speaker and timestamp, then rename speakers by their detected label and save.
Rename speakers
The diarization pipeline assigns raw labels —SPEAKER_00, SPEAKER_01, and so on. The Speakers panel lists each label with a text field where you type the real name.
1
Identify the speakers
Read the first few segments in the Transcript panel to match voices to labels.
2
Type the names
Click into each speaker field in the Speakers panel and type the person’s name. The placeholder reads “Type names to replace the raw labels”. You can also keep a raw label by choosing “(keep raw label)”.
3
Names update live
As you type, every occurrence of that speaker label in the Transcript panel updates in real time.
If speaker detection was disabled for this recording, all segments share one label. Type a single name to label the whole transcript (“Speaker detection was disabled for this recording, so all segments share one speaker.”).
The transcript panel
The Transcript panel shows every segment with its start and end timestamps and the speaker label. Segments are ordered chronologically. Scrolling through a long transcript, you can click any segment to jump the audio playback to that point (audio upload only).Audio timeline
For transcripts created from uploaded audio files, an audio playback bar appears at the bottom of the review sheet. It shows the waveform or a simple progress track, a play/pause button, and the current playback position. Click anywhere on the timeline to seek.Audio preview is not available for meeting recordings (“Audio preview isn’t available for meeting recordings.”).
Post-processing results
If post-processing prompts were attached to the job (meeting recordings only), a Post-processing section appears below the Transcript panel. Each configured prompt appears as a card:- Result text — the prompt ran and returned output.
- Re-run — the prompt ran previously; click to run it again (for example, after renaming speakers).
- Run — the prompt has never run (its auto-run was lost); click Run to start it now.
- Running… — a run is in progress or was claimed within the last 30 minutes.
- Failed — the last run produced an error; click Re-run to retry.
Prompts configured for a job but whose auto-run did not complete appear as Run cards so you can start them manually from the review sheet.
Save to Knowledge
When you are happy with the speaker names, click Save changes. IMP saves the transcript as an item in the built-in transcriptions knowledge context.Saving a transcript makes it available across IMP surfaces. You can pin it to a project’s Files tab or add it to a chat session via the Add knowledge menu. This cross-surface flow connects the review sheet to the broader Knowledge system — see Knowledge overview for how knowledge items work across the platform.
Delete a saved entry
To remove a saved transcript from the list, click the trash icon on its row. A confirmation dialog titled “Delete saved transcription?” appears. By default it removes only the list entry (“This removes the entry from this page. The saved transcript stays in the knowledge base unless you also select it below.”). Check Also delete the saved transcript from the knowledge base to remove the knowledge item at the same time. Click Delete to confirm.Discard a transcript
Click Discard in the review sheet to cancel the job. The confirmation dialog warns: “The audio stays in your files; the transcription job will be cancelled.” The uploaded audio file is kept in your file storage.Next steps
Record a meeting
Send a bot to join and record a live meeting.
Pin knowledge in chat
Add a saved transcript to a chat or project.