The audio and video API for speakers, transcripts, stems and captions
Split recordings by speaker, transcribe with speaker names, remove vocals, split stems, clean up speech, burn in captions, cut silences, translate and change voices, all through one REST API with signed webhooks.
Guides by tool
- Speaker diarization API: Speaker diarization API that returns real audio files: one time-aligned track per speaker, speaker names, and a transcript in SRT, VTT, TXT or JSON.
- Transcription API: Transcription API that labels every line with the speaker’s name. Get JSON, TXT, Markdown, SRT or VTT from one REST call, for files up to 4 hours long.
- Vocal remover and stem API: Vocal remover API and stem splitter API: split a song into vocals and instrumental, or six stems (vocals, bass, drums, piano, guitar, other) over REST.
- Auto caption API: Auto caption API that burns captions into your video in time with the speech, plus a silence remover that cuts pauses and filler words. One REST call each.
- Video translation API: Video translation API: translate the speech in a video or podcast into 21 languages, read aloud by an AI voice, with the translated text. Voice changer too.
- Audio MCP server: Connect the SpeakerSplit MCP server to Claude, ChatGPT, Cursor or Claude Code to split speakers, transcribe, remove vocals, caption and translate in chat.
- SpeakerSplit in Claude: Add the SpeakerSplit connector to Claude, then ask it to split speakers, transcribe, remove vocals, add captions or translate your recordings in plain words.
- SpeakerSplit in ChatGPT: Use SpeakerSplit in ChatGPT developer mode: attach a recording to the chat and ask it to split speakers, transcribe, remove vocals, caption or translate it.