Extract vocals, instruments, drums and bass from any audio file using state-of-the-art AI. Processed on your own server — your files stay private.
Drop your audio file here
or click to choose a tool below
MP3 · WAV · FLAC · OGG · M4A · AAC
Capabilities
Multiple AI models, each optimized for a specific separation task.
Cleanly separate lead vocals from the instrumental — perfect for karaoke, remixing or vocal analysis.
Get a pure drum track for remixes, practice sessions or beat analysis.
Isolate the bassline for transcriptions, learning or creative remixing.
Remove all vocals instantly — create karaoke tracks from any song in seconds.
Combine multiple AI models and average their outputs for the highest possible quality.
Your files never leave your own server. No accounts, no cloud, no data retention.
How it works
No signup, no waiting, no uploading to third-party servers.
MP3, WAV, FLAC, OGG, M4A or AAC — drag & drop or click to select. Up to 100 MB.
Choose from UVR MDX-Net, Kim Vocal 2, Demucs 6s, or a multi-model ensemble for best quality.
Listen to each stem directly in the browser, then download as WAV. Files are auto-deleted after use.
Tools
Each tool uses a different AI model, optimized for its specific task.
Extract vocals, instrumentals, drums, bass or karaoke using UVR MDX-Net models with optional ensembling.
Start separating →Split any instrumental into drums, bass, guitar, piano and other using Demucs 6-Stem AI.
Split instruments →Detect the musical key, BPM and full chord progression with an interactive timeline.
Analyze audio →