AI vocal remover
Remove the singer from a song and keep the full band. Download the instrumental for karaoke or covers, or the isolated vocals for remixes, all without uploading your music.
1. Choose a song
2. Separate
- The first run downloads the AI model (174 MB) once; your browser keeps it for next time. Your song is never uploaded.
3. Listen and download
The instrumental and the vocals appear here when the separation is done.
How it works
- Pick a song from your device.
- Press Remove vocals: an AI model separates the voice from the music, on your device.
- Listen to the instrumental and the vocals, then download the instrumental, the vocals, or both as a ZIP.
Make a karaoke track from any song
The vocal remover splits a song into the instrumental, which is everything except the lead voice, and the vocals. Because the AI model recognises the voice itself, the bass, drums, and guitars stay in the instrumental instead of being cut out along with the singer.
Use the instrumental for karaoke nights, covers and practice, backing tracks for your own singing, or a background track for a video. Download it with one click as WAV or MP3, or take both tracks in a ZIP. The vocals are there too, ready for remixes or to check a lyric.
Tips for a cleaner instrumental
A lossless file (WAV or FLAC) or a high-bitrate MP3 gives the best result. Trim the song to the part you need first to save processing time, and listen to the instrumental before you download it: solo it, and watch the vocals meter to see whether anything of the voice is left.
If you want more control, the stem splitter separates drums, bass, and other instruments as well, and the acapella extractor gives you the vocals as the main download. Planning to sing along? The BPM and key finder tells you the key to sing in.
Related:Acapella ExtractorStem SplitterAudio TrimmerBPM & Key Finder
Frequently asked questions
How is this different from old vocal removers?
Older tools cancel out whatever is in the centre of the stereo mix, which also removes bass and drums and fails on reverb. This tool uses an AI model that recognises the voice itself, so the instrumental keeps the full band.
Can I also get just the vocals?
Yes. You get two tracks: the instrumental, which is everything except the voice, and the isolated vocals. Solo either one to listen, and download them separately or together. The acapella extractor is built for the opposite job and puts the vocals first.
Will the instrumental sound exactly like the original band?
Very close on most songs. The AI model sometimes leaves a faint trace of the vocals (often reverb or backing vocals) or takes a little of the music with the voice, especially on dense or heavily processed mixes.
Is my song uploaded to a server?
No. The separation runs on your own device, inside your browser. The only download is the AI model itself; your audio never leaves your computer or phone.
How big is the AI model download?
The AI model is about 174 MB. It is downloaded once from SearchSound.cloud the first time you separate a song, then kept in your browser’s cache, so later songs start right away. Clearing your browser data removes it.
How long does separation take?
On a typical laptop it takes about 4 to 5 minutes per minute of music, because the model runs on a single processor core. Browsers with WebGPU (recent Chrome or Edge on a computer with a capable graphics card) can be faster. Phones are much slower and may run out of memory, so a computer works best. Songs can be up to 10 minutes long.
Why does it need so much memory?
The AI model is a large neural network and needs about 3 GB of memory while it works. If the tab crashes, close other tabs and apps, or trim the song first.
Can I use the separated stems in my own releases?
Separating a song does not change who owns it. Use stems of copyrighted songs for personal practice, study, and private remixes, and get permission from the rights holders before publishing or selling anything made from them.
Is this tool free?
Yes. It is completely free, with no account, no watermark, and no limit on how many files you process.
Which file formats are supported?
You can open MP3, WAV, FLAC, M4A, AAC, and OGG files, and download each stem as WAV or MP3, or all of them together as a ZIP file.