Speaker Separation
Turn a recording into a timestamped transcript with each speaker labeled. Rename speakers, download isolated audio stems, or export txt and srt.
Guest runs are free and rate-limited · No account needed
What Speaker Separation can do
- Identify and separate individual speakers
- Rename speakers; names carry into the transcript and downloads
- Download isolated per-speaker audio stems
- Timestamped transcripts you can copy or export as txt or srt
- Supports MP3, MP4, WAV audio files
- Works with meetings, interviews, and podcasts
Who uses Speaker Separation
Journalists
Separate interview speakers automatically for fast transcription.
Podcast editors
Download isolated speaker stems and a labeled transcript for easier post-production.
Legal teams
Attribute spoken words to specific participants in recorded hearings.
Frequently asked questions
Can AI transcribe multiple speakers?
Yes. Upload a recording and Speaker Separation labels each voice in a timestamped transcript. You can then rename those labels, download isolated audio stems, or export the transcript as txt or srt.
What's the difference between diarization and isolated stems?
Diarization labels who said what in the transcript. Isolated stems are separate audio files for each speaker, cut from the original recording. One run gives you both.
Can I rename Speaker A and Speaker B?
Yes. After the job finishes, rename speakers to real names. Those names carry into the transcript, timeline, and downloads.
What can I export?
Copy the labeled transcript, or download it as txt or srt. You can also download each speaker as its own audio file, or a zip of everyone.
Ready to try Speaker Separation?
Try a guest run now. Sign up later if you want to keep your work and get credits.
Try Speaker Separation without signing up