How to use this tool
- Drag and drop your audio or video file into the box, or click "Choose a file".
- For audio, pick the output format. Video is saved as MP4 and the picture stays as it is.
- Click "Remove music" and wait. The first time we load the separation model (about 172 MB) and store it in your browser.
- Listen to the result, then download it. Review it before relying on it, because automatic separation is not always perfect.
Frequently asked questions
Is my file uploaded to a server?
No. Separation happens inside your browser on your device, and we even host the model ourselves. You can use it for private recordings without them leaving your device.
What remains in the output file?
Only the human voice (speech or singing). Music and instruments are removed. For video, the picture stays untouched and only the sound changes.
How accurate is the separation?
Good in most cases, especially speech over clear background music. But faint traces of music may remain or the voice may be slightly affected, and it is weaker with overlapping voices, old recordings or a lot of noise.
Why does it take time?
Because the model runs on your device's processor. On a four-core machine it takes roughly one and a half times the clip's length. Keep the page open; the stronger your device, the faster it goes.
How long a file can I use?
Up to 15 minutes on a computer and 5 minutes on a phone. Longer files need a lot of memory, so split them into parts.
Can I keep the music and remove the voice instead?
This tool removes the music and keeps the voice. Removing the voice and keeping the music would be a separate tool we may add later.
Can I use copyrighted songs?
You are responsible for the rights to the files you process. Use the tool for your own content, or content you are allowed to work with.