Turn an interview or meeting into a transcript with speaker names
Whisper transcribes the recording on your device, then a speaker model finds who speaks when: each line is split where the voice changes, given a voice print, and the prints are grouped into Speaker 1, Speaker 2 and so on, which you rename. The result is a transcript with names and times, ready as a Word document.
Every step runs in your browser; your files are not uploaded.
Steps
Transcribe it
Choose the recording (audio or video). Whisper writes a timed transcript, with the time of every word, and sends it on with the recording.
Leave the number of speakers on Find out, or pick it if you know it. Each line is split where the speaker changes for at least 0.8 s, then grouped by voice. Type the speakers' names, check the preview, and send it on.
Workflow step
Save as Word
Each turn becomes a paragraph starting with the speaker's name and the time.
