Demucs

Meta's AI music source separation model for isolating vocals, drums, bass, and more

AI MusicFree

Demucs is Meta AI's open-source music source separation model that splits a mixed audio track into individual stems including vocals, drums, bass, and other instruments with state-of-the-art quality. Its hybrid transformer architecture produces clean separations that minimize bleeding between tracks. Music producers, remixers, audio engineers, and karaoke app developers use Demucs to extract stems for remixing, create instrumental versions, analyze individual instruments, and restore old recordings where original stems are unavailable.

Key Features

• ✓Stem separation
• ✓Vocals extraction
• ✓Drum isolation
• ✓High quality
• ✓Open-source