Vocal remover and stem splitter API
Vocal remover and stem separation API
Separate the voice from the music, or split a whole song into six stems, with one request. It is the same separation the SpeakerSplit vocal remover and stem splitter run on the website, ready for your own app or pipeline. A third tool removes background noise from speech.
Get an API key ยท Read the API docs
Pick the split
| tool | What you get |
|---|---|
| vocal_split | A vocals track and an instrumental track. |
| stem_split | Six stems: vocals, bass, drums, piano, guitar and other. |
| audio_enhance | One cleaned-up WAV, with background noise removed from speech. |
Split a song into stems
The finished job lists its files in outputs. Each one has a name, url, filename and expires_at, and the links work for 30 days.
curl https://speakersplit.io/v1/jobs \
-H "Authorization: Bearer $SPEAKERSPLIT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"tool": "stem_split",
"audio_url": "https://cdn.example.com/tracks/demo-mix.wav",
"webhook_url": "https://example.com/hooks/speakersplit"
}'Clean up speech, too
audio_enhance removes background noise from speech. Leave enhance_mode at "denoise" to keep the voice as it was recorded, or set "enhance" to also restore muffled speech. Enhance can change how the voice sounds.
curl https://speakersplit.io/v1/jobs \
-H "Authorization: Bearer $SPEAKERSPLIT_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "tool": "audio_enhance", "audio_url": "https://cdn.example.com/raw/interview.mp3", "enhance_mode": "denoise" }'Files it takes
Audio like MP3, WAV, M4A, FLAC and OGG, or video like MP4, MOV and WEBM, up to 90 minutes long. Video files cost the video rate.
Limits and safety
- 120 requests a minute per API key, and up to 5 jobs running at once per account.
- Upload local files up to 2 GB with POST /v1/uploads, or pass a public https link as audio_url.
- Send an Idempotency-Key header and a retried request never starts (or charges) a second job.
- Unknown parameters are rejected with a 400, so a typo never runs a job with the wrong settings.
- Result links work for 30 days.
What it costs
| Length | Vocals or clean-up, audio | Vocals or clean-up, video | Stems, audio | Stems, video |
|---|---|---|---|---|
| Under 5 minutes | 1 | 2 | 2 | 3 |
| 5 to 30 minutes | 2 | 4 | 3 | 6 |
| 30 to 60 minutes | 4 | 6 | 6 | 9 |
| 60 to 90 minutes | 6 | 8 | 9 | 12 |
Stems cost 1.5 times the vocal price, rounded up. Credits are the same ones you use on the website. Plans and credit options are on the pricing page.
Common questions
What is the difference between vocal_split and stem_split?
vocal_split gives you two files: the vocals and everything else (the instrumental). stem_split gives you six: vocals, bass, drums, piano, guitar and other.
Can a vocal remover separate two people talking?
No. A vocal remover separates voices from music, so every voice ends up on the same vocals track. To give each speaker their own track, use the speaker diarization API (tool speaker_split).
How long can a song be?
Up to 90 minutes per file, audio or video.
How much does the vocal remover API cost?
For audio, 1 credit under 5 minutes, 2 under 30, 4 under 60 and 6 up to 90 minutes. Video costs 2, 4, 6 and 8. Stems cost 1.5 times that, rounded up.
Is it the same quality as the SpeakerSplit website?
Yes. The API runs the same separation as the vocal remover and stem splitter on the website, and charges the same credits.