← Back

How it works Vocal remover FAQ

Karaoke
Maker

Turn any song into a backing track — original key, original tempo, vocal gone

MP3, WAV or FLAC · up to 20 MB and 5 minutes · 25 credits a track

How to make a karaoke track

  1. Upload the song. Any MP3, WAV or FLAC up to 20 MB and five minutes.
  2. The AI takes the vocal out. A model trained on isolated stems separates the sung voice from everything else. Nothing is transposed or stretched, so the key, tempo and arrangement stay exactly as you know them.
  3. Set your guide vocal. This page opens with the vocal fader already down. Leave it there for a full karaoke track, or lift it to about 15–20% to keep the original singer quietly underneath while you learn the song.
  4. Save the backing track. MP3 for a phone or a PA, WAV if it is going into a DAW or a video.

What it is good for

  • Rehearsing — a backing track for a song there is no official karaoke version of.
  • Singing lessons — a guide vocal you can fade down as you learn it.
  • Live sets — a track to perform over, at the key the record is in.
  • Covers and videos — the instrumental to sing over on camera.
  • Parties — a karaoke night built from your own playlist.
  • Auditions — the accompaniment on its own, with no voice competing.

Frequently asked questions

Does the karaoke track stay in the original key?
Yes. Nothing is transposed or time-stretched — the music track is the original recording with the vocal taken out, so key, tempo and arrangement are untouched. If you need it in a different key, run the result through the Key & Tempo Shifter afterwards; it changes one without dragging the other with it.
Can I keep a quiet guide vocal in?
That is exactly what the vocal fader is for. At zero you get a full karaoke track; at 15–20% the original singer sits under you as a guide, which is how most people rehearse a song they do not know by heart yet. You hear the balance before you save it.
Will backing vocals and ad-libs come out too?
Mostly. The model separates sung voice in general rather than just the lead. Heavily processed backing vocals, vocal chops used as an instrument and talkbox parts sometimes stay with the music — musically, that is what they are.
Is it good enough to perform with?
For a party, a rehearsal or a video, comfortably. On a loud PA you may hear faint traces of the original vocal in the busiest sections — that is true of every AI separator, including paid ones. Dense, heavily compressed masters are the hardest case; sparse arrangements come out almost perfectly.
Can I use a karaoke track I made in public?
The tool gives you the audio, not the rights. Public performance and anything commercial still needs the usual licensing for the underlying song, the same as any other karaoke track. Private practice and your own material are the safe ground.
What if I want the vocal instead of the music?
Push the vocal fader up and the music fader down, and Save gives you the acapella. The vocal remover and isolation page is the same tool with both faders starting up.

Related tools