Skip to content
← Help

Working with the transcript

Every meeting has a full transcript you can read, navigate, and play back alongside the recording.

Open the transcript

Open a meeting and select the transcript tab. The transcript is laid out as speaker turns, each labelled with the speaker and, where available, a timestamp. Your own turns are styled separately because Daisy knows which speaker is you, and other speakers are tinted so you can scan who said what. A Room or Remote badge marks whether a speaker was in the room with you or on the far side of a call.

If the meeting is still being processed, this tab shows the live transcript under a Live transcript — finalizing… card, and updates on its own once the cleaned, diarized version is ready. The words themselves don't change between the two — Daisy transcribes once, while you record, and keeps that text. What arrives with the finished version is the cleanup around them: speaker names, echo removal, and filler trimming.

You may see a standalone line like [possible missed speech — meeting audio was faint for 4s]. That's a note: Daisy heard something there that didn't decode into words, usually very faint audio. The audio itself is on disk and the player will scrub to it. You can turn these markers off under Settings → Providers → Advanced → Transcription tuning.

Jump to a moment in the recording

When a recording is available, an audio player with standard play, pause, and scrub controls sits above the tabs.

  • Click any turn's timestamp to seek the player to that moment and start playing.
  • Click a chapter in the chapters panel to jump to where that topic begins.

Label who's speaking

Speaker names come from the participants tab. Open that tab to group voices, name speakers, and play a sample of each one. See Identifying who said what for the full workflow.

Copy the transcript

The Copy all button on the transcript tab's bar copies the whole transcript to your clipboard. If no AI summarizer is configured, Daisy prepends a prompt so you can paste it straight into ChatGPT, Claude, or another chat.

Regenerate the transcript

To rebuild the transcript — for example after importing better audio — open the Action menu and choose Regen. Transcript. This re-transcribes on your machine, re-runs de-duplication, and re-groups the speakers. Daisy asks you to confirm first, since it overwrites the existing transcript.

Polish the transcript

Action → Polish Transcript sends the transcript to your configured AI provider to clean up punctuation and casing and redact obvious secrets, then overwrites the transcript text. It needs an AI provider and does nothing on an empty transcript.

Trim or split a recording that ran long

If the recorder kept going after the meeting ended — or two meetings ran into one file — you can cut the session at any point in the transcript.

Open the Transcript tab, scroll to the line where the meeting should end or split, then right-click that line:

  • Split into new meeting — everything after that line moves into a new recording titled Split recording, keeping its own date and time. Tags and title don't follow it. Nothing is lost.
  • Delete from here — everything after that line is removed: recording, transcript, and summary. This can't be undone.

Either way, the line you right-clicked and everything before it is kept, and Daisy asks you to confirm first — naming the timestamp and quoting the line so you can be sure of the cut point.

Both are also in the Action menu (⤨ Split into new meeting, ✂ Delete from a point…), where they point you back to the transcript — the cut point has to be picked on a line.