Vocal remover: split a song into vocals and instrumental
Split any song into vocals and instrumental (karaoke) with AI — free, right in your browser.
🔒 Files stay on your device✓ Free, no sign-up
Free and private: the song is not uploaded anywhere. The first time the AI model (~64 MB) is downloaded, then it starts at once. Works best in Chrome or Edge on a computer.
Your browser has no WebGPU, so the AI will run on the processor: a 3-minute song can take 15–20 minutes. Chrome or Edge on a computer is much faster.
AI model: Kim Vocal 2 (UVR MDX-Net) by KimberleyJensen · Ultimate Vocal Remover.
How it works
Drop a song. The AI separates the voice from the music and gives you two WAV files: the a cappella vocal and the instrumental (karaoke / minus track). Everything happens on your device; the first time the browser downloads the AI model (~64 MB).
How long does it take?
With a modern graphics card (Chrome or Edge on a computer) about as long as the song itself. Without WebGPU it works on the processor and is several times slower.
Is the quality like in paid services?
It uses Kim Vocal 2 — one of the best-known vocal models from Ultimate Vocal Remover. Reverb and backing vocals may stay partly in the vocal track; dense mixes are the hardest.
Are my files uploaded anywhere?
No. Everything is processed by your own browser. Your music is not sent to our server or anyone else, so the tool is private and works even with very large files.