What is Suno Stems?
Suno Stems splits a finished Suno song into separate tracks so you can work on each part on its own. It acts on a clip from an earlier Suno generation, identified by its clip_id, and there are two models behind this page. suno-ai/stems does a two-way split into vocals and instrumental, which covers karaoke tracks, a cappella exports and simple remixes. suno-ai/stems-all goes further and returns up to twelve stems: Vocals, Backing Vocals, Drums, Bass, Guitar, Keyboard, Percussion, Strings, Synth, FX, Brass and Woodwinds. The request body is the same for both, a required clip_id plus an optional title, so switching between them is a one-word change in the model field.
Key features of Suno Stems
Three things set this endpoint apart from the other Suno operations.
Two-way split for vocals and instrumental
suno-ai/stems returns a vocal track and an instrumental track. It is the simple option when you only need to remove or isolate the singer.
Full stem set with stems-all
suno-ai/stems-all separates the mix into Vocals, Backing Vocals, Drums, Bass, Guitar, Keyboard, Percussion, Strings, Synth, FX, Brass and Woodwinds, depending on what the song contains.
One small request body
Both models take only clip_id (required) and title (optional). There are no model versions or prompts to set, which makes stem extraction easy to add as an automatic step after generation.
Best for
Karaoke tracks
Run the two-way split and serve the instrumental stem as the backing track.
Remixes
Pull drums, bass and vocals apart with stems-all and rebuild the song in a DAW.
Clean vocal exports
Get the lead vocal on its own for tuning, doubling or use in another song.
Mixing a generated song properly
Balance, EQ and compress each instrument separately instead of working on a single stereo file.
Sample packs
Collect drum, synth and FX stems from generated songs to build loops and one-shots.
Use cases
A music generation app can show a "split stems" option on each finished song that sends its clip_id to suno-ai/stems and offers the vocal and instrumental files for download. A remix platform can run suno-ai/stems-all on every new track so users can mute or swap individual instruments in the browser. A karaoke service can generate songs and strip the vocals in the same pipeline, with the stem request fired as soon as the generation task completes. To export the whole mix as one WAV file instead of separate parts, use the Suno WAV Download API.
API examples
Both stem models are documented at docs.unifically.com/models/audio/suno/stems.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "suno-ai/stems-all",
"input": {
"clip_id": "1901ec96-72c6-4c26-8e56-91028ca6070d",
"title": "All My Stems"
}
}'
Swap the model to suno-ai/stems for the two-way split. Either call returns a task_id; poll GET /v1/tasks/<task_id> for the stem files or set a callback_url to receive them when they are ready.
FAQs
People also ask
It separates a Suno song into individual tracks. Send the song's clip_id and the job returns stem files you can mix, mute or edit on their own.
suno-ai/stems gives you two tracks, vocals and instrumental. suno-ai/stems-all splits the song into the full stem set: Vocals, Backing Vocals, Drums, Bass, Guitar, Keyboard, Percussion, Strings, Synth, FX, Brass and Woodwinds.
Not always. A stem can only come back if that instrument is in the mix, so a piano ballad will not produce a useful guitar or brass track. Expect the set to follow what the song actually contains.
This endpoint takes a clip_id only, with no audio_url field, so the input has to be a clip from an earlier Suno generation.
Just clip_id, which is required. An optional title names the extraction so it is easier to find in your task list.
