separate_music_vocals splits a track you generated in-app into its vocal and instrumental stems — useful for building a karaoke version, remixing, or pulling just the instrumental bed for a video that already has its own voiceover.
It's specific to Versely-generated tracks: you need the source track's taskId and audioId from a prior generate_music or extend_music result, not just any uploaded file (for that, isolate-vocals-from-a-track is the general-purpose option).
Powered by
Split a previously generated Suno track into vocal/instrumental stems. Requires the source track's taskId and audioId (from a prior generate_music / extend_music result).
What to tell the agent
Versely's agent maps this plain-English request directly onto separate_music_vocals. You don't need to know the parameter names — just describe what you want.
“Split my last generated track into vocal and instrumental stems.”
How it works
1. Find the source track's IDs
Pull taskId and audioId from the prior generate_music or extend_music result.
2. Choose the stem type
type selects which stem(s) you want back.
3. Run the split
Versely separates the track into vocal and instrumental stems.
4. Use the stems
Use the instrumental as background music (see add-music-to-video), or the isolated vocal on its own.
What it costs
Billed per split, on top of the original track's generation cost. Confirm the exact price with estimate_cost.
The formula behind that number: What does a finished 30-second AI ad cost end to end? — sums four unrelated meters, so no single rate predicts the total.
Limits & things to know
- Only works on tracks generated in-app via generate_music/extend_music, identified by taskId + audioId — it's not for arbitrary uploaded audio files.
- For splitting vocals out of an arbitrary recording (not a Versely-generated track), use isolate_audio instead.
Who uses this
- Building a karaoke-style instrumental version of a generated track
- Pulling just the instrumental bed to use as background music
- Remixing or re-arranging a generated song's components
Frequently asked questions
Can I split any music file into stems?+
No — separate_music_vocals only works on tracks you generated in-app via generate_music or extend_music, referenced by taskId and audioId. For an arbitrary uploaded recording, use isolate_audio instead.
What do I get back?+
Separate vocal and instrumental stems from the source track, selectable via the type parameter.
Related Versely tools
Related editing jobs
Split a Song Into Stems inside Versely
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.