SplitFire AI

SplitFire AI splits a song into stems: vocals, drums, bass and more. The
separation runs on your computer, so your audio is not uploaded anywhere.

Pick an audio file, choose a model, and press Separate. There are three
HT-Demucs models: a fast 4-stem model, a fine-tuned 4-stem model and a
6-stem model that adds guitar and piano. Each model downloads the first
time you use it and is cached after that.

Stems are saved as WAV files, so you can open them in any editor or DAW.
History keeps your past separations, with playback for each stem.

You don't need an account to separate audio.