What the Vocal Remover does
It takes one mixed track and separates it into two results: the vocal and the instrumental. That is useful when you want a backing track for karaoke, a bed for a video, or a vocal to study and edit. It is a separation task, not a remix: the arrangement stays the same, and the parts should line up when you play them together.
It is also not a transcription tool. You do not get MIDI, sheet music, or per-instrument stems. You get two audio files.
The important limit: workspace tracks only
The Vocal Remover picks its source from tracks already in your account. It is not a drag-and-drop uploader for arbitrary commercial songs. If you want to separate a track you generated here, keep it in your workspace and select it in the tool. If you want a vocal-free arrangement from a text brief instead, use Instrumental Mode.
That distinction matters for both quality and rights. A track you generated here has a known provider and a known usage path. A random commercial recording does not.
How to run it
- Open the AI Music Generator with the
mode=vocal-removertab selected. - Pick a source track from the list. Only tracks with a usable audio source appear.
- Read the credit cost next to the selected model.
- Submit the task. The workbench polls for the result.
- Preview both parts. Download the one you need, or both.
If the source list is empty, generate a track first. The tool cannot split what it does not have.
What good separation sounds like
- The vocal part should contain the lead voice, with some bleed from the instruments. Some bleed is normal.
- The instrumental should keep the drums, bass, and harmony. If the vocal is still audible, the source may be too dense or the model may need a different lane.
- The two parts should stay in sync. If they drift, the separation failed; regenerate rather than trying to align them by hand.
Do not expect studio-grade stems. Expect a usable backing track and a usable vocal reference.
When to use which tool
| Goal | Tool |
|---|---|
| A new song with no vocals | Instrumental Mode |
| Split an existing generated track | Vocal Remover |
| Change a section of an existing track | Replace Section |
| Make a track longer | Extend |
| Add a vocal to an instrumental | Add Vocals |
Matching the tool to the goal is the cheapest way to work: the wrong tool produces a result you cannot use, and the task is still billed.
Cost
Vocal removal is a billed task, and the credit cost is shown before submission. It uses the selected model’s rate, so the cheapest lane is not always the right one: if the source is dense, a better lane may produce a cleaner split. Budget one or two attempts.
The $9.90 Starter pack contains 100 credits. At a 10-credit lane, that is about ten tasks; a single split leaves most of the pack for the rest of the project. Credits do not expire, and the pack does not renew. The pricing page lists every lane and its credit cost.
Rights and reuse
Separating a track does not create new rights. The vocal and the instrumental remain subject to the same Terms, provider restrictions, and applicable law as the source. If you publish the instrumental, treat it as a paid output, not as a new work with a fresh clearance.
Do not use the tool on recordings you do not have the rights to process. The workspace restriction is there for a reason.
Honest limits
- Workspace tracks only. No arbitrary uploads.
- Two parts, not stems. No per-instrument isolation.
- No MIDI or notation. Separation is audio-only.
- No free credits. New accounts start at zero; the entry pack is $4.99.
Open the workbench, pick a track you already generated, and read the cost before you run the split.