SpeakerSplit for ChatGPT

Split a recording into one audio file per speaker, in ChatGPT

Attach a podcast, interview or meeting recording to ChatGPT and ask it to split the speakers. SpeakerSplit puts each person on their own track, and ChatGPT gives you the download links.

Set it up · Everything it does in ChatGPT

Can ChatGPT separate speakers on its own?

ChatGPT doesn’t come with a tool that cuts a recording into a separate audio file for each voice. The SpeakerSplit connector adds one: it works out who is speaking when, and gives each person their own track.

Set it up in ChatGPT

SpeakerSplit connects to ChatGPT through developer mode.

  1. In ChatGPT, turn on developer mode.
  2. Add a connector with the address https://speakersplit.io/mcp
  3. Sign in to SpeakerSplit and press "Allow".
  4. In a chat, attach your recording and say what you want done.

What to type

  • "Split this interview into one track per speaker."
  • "Split this podcast. The host is Jane Doe and the guest is Sam Lee."
  • "Split this panel recording. There are four speakers. Give me WAV files."

What you get back

  • One audio file per speaker, as MP3 or WAV. Each runs the full length of the recording and is silent while that person isn’t talking, so the tracks line up in an editor.
  • Names instead of Speaker 1 and Speaker 2. Say who is in the recording, up to 10 people, and the tracks and transcript use their names.
  • A transcript with every line labelled, which you can also ask for as SRT or VTT subtitles.

What it costs

Recording lengthCredits
Under 5 minutes1
5 to 30 minutes2
30 to 60 minutes4
60 to 90 minutes6
Each started 30 minutes after that, up to 4 hours+2

Credits are the same ones you use on the website. Plans and credit options are on the pricing page.

See plans and credits

Common questions

Can ChatGPT separate speakers in an audio file?

With the SpeakerSplit connector, yes. Attach the recording and ask ChatGPT to split it by speaker. You get one audio file per speaker, their names and a transcript.

Do the tracks line up with each other?

Yes. Every track runs the full length of the recording and is silent while that person isn’t talking, so you can drop them straight into an editor.

How many speakers can it split?

Up to 10 named speakers. Saying how many people speak, if you know, helps.

Related