Product disclosure statement
AI stem splitting
Read this before you split a track. It explains what you get, how long it takes and what it costs.
1. What the service does
You upload a finished mix and an AI model separates it into 6 parts: Vocals, Drums, Bass, Guitar, Keys, Other. The results are MP3 files you can preview, keep or discard before they're added to a song. Separation is an estimate made by the model — it is not the original multitrack recording.
2. Accepted files
Accepted formats: MP3, WAV, AIF, AIFF, FLAC, M4A, AAC, OGG, OPUS, WMA. The largest file accepted is 200 MB.
Lossless files (WAV, AIFF, FLAC) give the cleanest results but are much larger. A compressed copy (a high-quality MP3) of the same song usually costs less and finishes faster.
3. Credit charges
Each track uses prepaid AI credit. The amount depends on the size of the file you upload:
- Standard — up to 60 MB: 1× the standard track price
- Large — over 60 MB, up to 120 MB: 2× the standard track price
- Extra large — over 120 MB, up to 200 MB: 3× the standard track price
The exact amount is shown before you confirm. Nothing is charged until you press "Accept and split".
4. How long it takes
Most standard songs finish in 1 to 5 minutes. Long tracks and large lossless files can take 10 minutes or more. At busy times the job may wait in a queue before it starts.
The time shown is an estimate only. You can leave the page — the job keeps running and the stems will be waiting for review when you come back.
5. Refunds
- If a split fails or is cancelled by the service, the credit goes straight back to your balance.
- Once stems are delivered, the credit is spent — even if you discard them — because the processing has been done.
- Separation quality varies with the recording; it isn't grounds for a refund on its own.
6. Review and storage
Finished stems wait for your review for 7 days. Confirmed stems are saved to the song you choose and count toward your plan's track slots.
7. Your rights to the music
Only upload recordings you own or have permission to use. You're responsible for how the separated stems are used, including in live performance.
8. Privacy
Uploaded audio is sent to our processing partner only to perform the split. Files stay private to your act, are never shown on public pages, and aren't used by us to train AI models.