Model · Music and sound
Demucs
A free, open model that splits a song into separate drums, bass, vocals and other parts (a six-part version adds guitar and piano). You run it yourself from the command line; it powers many stem-splitting apps.
At a glance
- Open or closed
- Fully openWeights and code downloadable under a standard open licenceCode and pretrained weights downloadable under MIT.source · checked 12 Oct 2026 · checked on the maker's own page
- Licence
- MIT (Permissive (e.g. Apache 2.0, MIT))source · checked 12 Oct 2026 · checked on the maker's own page
- Can I sell what I make?
- YesMIT licence allows commercial use. Rights in the song you split still belong to its owners.source · checked 12 Oct 2026 · checked on the maker's own page
- Cost
- Free option: Free to download and run yourselfPlans below.
- How you use it
- On your own computerInstall with pip (Python 3.8+). 3GB GPU minimum is stated for Demucs generally.source · checked 12 Oct 2026 · checked on the maker's own page
- Skill needed
- Technicalsource · checked 12 Oct 2026 · checked on the maker's own page
- Made by
- Meta (Facebook Research)Published under the facebookresearch GitHub organisation; no company location stated.source · checked 12 Oct 2026 · checked, but the page was unclear or partial
- Status
- AvailableNo lifecycle state fits 'archived but still available'; kept as active.source · checked 12 Oct 2026 · checked on the maker's own page
What you can make with it
- Music and sound, separate a song into parts (main use)
audio to audio
source · checked 12 Oct 2026 · checked on the maker's own page
Plans and prices
No paid plans listed.
source · checked 12 Oct 2026 · checked on the maker's own page
Other options for “Music and sound, separate a song into parts”
AudioShake, Eleven Music, Filmora, Google Flow Music, Google Flow Music (formerly MusicFX), Google Flow Music (formerly Riffusion), Kits AI, LALAL.AI, LANDR, Loudly, Moises, MyEdit, Soundful, Soundverse, Suno, Ultimate Vocal Remover, Voice-Swap
About this entry
Facts last checked 12 Oct 2026 against the maker’s own pages (each fact links to its source). Entry status: draft. Plain-English summary written by Plainly.