AI Vocal Remover

Split vocals and music from any song.

  • AI
  • Free preview
  • No sign up

You may also like

See all tools

Learn more

Separate the voice from the music with AI

This vocal remover splits a song into two tracks: the vocals alone (an acapella) and everything else (an instrumental, for karaoke or a cover). It uses Demucs, an AI model trained on thousands of songs to recognize a voice among instruments, so it works on real, finished songs, unlike the old trick of cancelling the center of a stereo mix. You hear a free 30 second preview before paying anything, and the full tracks are WAV files the length of the whole song.

How to make an instrumental of a song

  1. Drop your song in the box, or click to choose it. The free preview is ready in a few seconds.
  2. Press Play and switch between Instrumental, Vocals only and Both, or set your own levels.
  3. Click Get the full tracks, pay €1.99 once, and download the instrumental, the vocals or your mix.

What to use it for

  • Karaoke: the instrumental of any song, even one that has no official karaoke version.
  • Singing practice and covers: keep the voice at a low level as a guide, then remove it completely.
  • Acapella for remixes and DJ sets: the isolated vocals, ready to put on another beat.
  • Learning an instrument: hear the band without the singer, or the words without the band.
  • Videos: background music without lyrics under a voice over.

Good to know

  • Songs up to 10 minutes. For a longer recording, cut it with the audio cutter first.
  • The full song is separated after the payment, in parts of 20 seconds: count about 5 seconds per part, so under a minute for most songs. Keep the tab open.
  • If something fails after the payment, "Try again" goes on with the same payment. If it keeps failing, contact us with your receipt and we refund you.
  • Only use songs you have the right to use: separating a track does not change who owns it.

Frequently asked questions

How do I remove the vocals from a song?

Drop the song (MP3, WAV, M4A, OGG or FLAC) in the box. The AI separates a free 30 second preview, taken from the loudest part of the song, where the voice usually is. Press Play, choose Instrumental to hear the song without its vocals, and if it sounds right, get the full tracks to download the whole song as WAV files.

Is the vocal remover free?

The preview is free: 30 seconds of the vocals and of the instrumental, played in the page, with no account. The full tracks of the song (vocals, instrumental and your own mix, as WAV files) cost €1.99, paid once for that song. There is no subscription, no credit and no minutes to buy in advance.

Which AI does it use?

Demucs (Hybrid Transformer Demucs, by Meta AI research), one of the best open models for music source separation. It was trained to split a song into voice, drums, bass and other instruments; we keep the voice on one side and add the rest together for the instrumental.

Is my song uploaded?

Yes, the AI runs on a server: the song is sent in short parts to our AI provider, Replicate, which separates each part and deletes it automatically within an hour. The song is never kept, published or used to train models. The parts are joined back together in your browser.

Will the instrumental sound perfect?

On most pop, rock and hip hop songs the voice is removed cleanly. Strong reverb, a choir, or a voice doubled by an instrument can leave a faint trace, and very old mono recordings separate less well. That is why you hear the preview first: if the 30 seconds sound good, the whole song will too.

Can I keep a little of the voice?

Yes. The Vocals and Music sliders set the level of each track. With the voice at 20 or 30%, you get a guide vocal for singing practice; with the music low, you hear the voice almost alone. Your mix plays live, and "Your mix" downloads it as a WAV file with the full tracks.

Related tools

  • Audio cutter: trim the instrumental, keep the chorus for a ringtone, add a fade out.
  • Video to GIF: turn a clip of your cover into a GIF to share.