SpeakerSplit for ChatGPT
Transcribe audio files in ChatGPT, with every speaker named
Connect SpeakerSplit to ChatGPT, attach a recording and ask for a transcript. Every line comes back labelled with who said it, and ChatGPT can then summarise it, pull out quotes or turn it into subtitles.
Set it up · Everything it does in ChatGPT
Can ChatGPT transcribe audio on its own?
It depends on your ChatGPT plan, and OpenAI changes it often. With the SpeakerSplit connector the result is the same every time: a transcript where each line carries the speaker’s name, subtitles timed to the speech, and recordings up to 4 hours long.
Set it up in ChatGPT
SpeakerSplit connects to ChatGPT through developer mode.
- In ChatGPT, turn on developer mode.
- Add a connector with the address https://speakersplit.io/mcp
- Sign in to SpeakerSplit and press "Allow".
- In a chat, attach your recording and say what you want done.
What to type
- "Transcribe this recording."
- "Transcribe this interview. The speakers are Jane Doe, the host, and Sam Lee, the guest."
- "Transcribe this lecture and give me SRT subtitles."
- "Transcribe this meeting, then list what each person agreed to do."
What you get back
- Every line labelled with who said it. Say who is in the recording, up to 10 people, and the labels use their names.
- The transcript as plain text, Markdown, JSON, or SRT and VTT subtitles timed to the speech.
- Recordings up to 4 hours long, audio or video.
What it costs
| Recording length | Credits |
|---|---|
| Under 5 minutes | 1 |
| 5 to 30 minutes | 2 |
| 30 to 60 minutes | 4 |
| 60 to 90 minutes | 6 |
| 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.
Common questions
Can ChatGPT transcribe an audio file?
With the SpeakerSplit connector, yes. Attach the file to the chat and ask for a transcript. Each line is labelled with the speaker, and you can ask for SRT or VTT subtitles.
Can ChatGPT tell who is speaking?
Yes, through the connector. Each line is labelled by speaker. Tell ChatGPT who is in the recording, up to 10 people, and it uses their names.
How long can the recording be?
Up to 4 hours. Under 5 minutes costs 1 credit; the table above shows the rest.
Do I need developer mode?
Yes. SpeakerSplit connects to ChatGPT through developer mode.