🎤 Creative Suite

Vocal Remover & Karaoke Maker

Strip the vocal out of a track for karaoke, or isolate it for practice. Runs in your browser — nothing uploaded.

Overview

Vocal Remover: karaoke tracks and isolated vocals in your browser

Most stereo mixes put the lead vocal dead centre — identical in both the left and right channels. That fact is the whole basis of centre-channel removal: subtract one channel from the other and anything perfectly centred cancels out, leaving the instruments that were panned to the sides. Reverse the operation and you get the opposite, an approximation of the vocal on its own.

This runs entirely in your browser. The audio is decoded, the maths is done on the raw samples, and a WAV comes back out — nothing is uploaded, there is no queue, and there is no length limit beyond your device's memory. It is worth being honest about what it can and cannot do: this is deterministic signal processing, not an AI stem separator, so it works beautifully on some tracks and imperfectly on others. The centred-content meter tells you which you have before you commit.

100% freeNothing uploadedNo sign-upInstant
How to use it
  1. 1Choose a song, or drop it onto the page.
  2. 2Check the centred-content meter — a high percentage means a clean result is likely.
  3. 3Pick Remove vocals for karaoke, or Isolate vocals to practise or study a performance.
  4. 4Adjust strength and the bass-keep control, then process and download the WAV.
Key features
  • Remove the vocal for a karaoke backing track, or isolate it on its own
  • A centred-content meter that predicts how well the track will separate
  • Variable strength, so you can reduce a vocal rather than fully cancel it
  • Bass-keep control that preserves the low end instead of gutting the kick and bass
  • Warns you when a file is mono, where the technique cannot work at all
  • Side-by-side playback of the original and the result, and WAV download

Frequently asked questions

How does removing vocals actually work?
In most mixes the lead vocal is panned to the centre, meaning it is identical in the left and right channels. Subtracting one channel from the other cancels anything identical in both and leaves whatever was panned to the sides. It is simple arithmetic on the samples, not a machine-learning model.
Why does some of the vocal still come through?
Because it is not perfectly centred. Reverb, delay and double-tracked harmonies are deliberately spread across the stereo field, so they survive the cancellation. Anything else centred — usually bass, kick and snare — gets removed along with the voice, which is what the bass-keep control compensates for.
Is this as good as an AI stem separator?
No, and it is worth being clear about that. AI separation models are trained to recognise what a voice sounds like and can pull one out of a mono recording. This technique only knows about stereo positioning. The trade-offs are that it is instant, needs no server, and works identically every time.
Is my music uploaded anywhere?
No. The file is decoded and processed in your own browser, and the result is built there too. Nothing is transmitted, which is also why there is no length limit or queue.
📖

Full guide

How Vocal Removal Works: Phase Cancellation, Stereo Fields and Its Limits

Why subtracting one channel from the other removes a vocal, what makes some songs work far better, and where AI separation differs.

Phase cancellationWhy mixes differAI vs arithmetic

Related Articles & Guides

BlogPDF Read Aloud: Text-to-Speech Guide for PDFsImagine finishing that 40 page research report during your morning commute, or absorbing a dense legal contract while cooking dinner. PDF read aloud technology ArticleHow Vocal Removal Works: Phase Cancellation, Stereo Fields and Its LimitsWhy subtracting one stereo channel from the other removes a centred vocal, what makes some songs separate far better, and where AI models differ.ArticleSubtractive Synthesis Explained: Oscillators, Filters and EnvelopesHow the classic synth signal path works, what each control actually changes about the sound, and how to build a patch from scratch.