MusicletYour everyday music toolkit.

Stem Splitter

Separate your song into vocals, drums, bass and other instruments.

Drop an audio file here

MP3, WAV, M4A, FLAC and more

How to split a song

Separate the mix

Choose a song to split it into vocals, drums, bass and other instruments.

Hear each part

Solo a track to listen closely, mute a part to practise, or adjust the levels together.

Take the stems with you

Save a single stem or all four in a ZIP. Choose WAV, or M4A when your browser supports it.

Are guitar and piano separate tracks?

No. This version produces four stems. Guitar, piano and other sounds share the Other track; some sound can bleed between stems.

What does my browser need?

Stem splitting needs WebGPU and enough memory for your song. Keep this tab open during processing. A short clip is a useful first test on a phone.

Why is there a model download?

Musiclet downloads the analysis files to run on your device. They are reused while your browser keeps its cache. Check progress in Models at the top of the page.

Which files can I use?

MP3, WAV, M4A and other audio formats your browser can decode. Support varies by browser and codec. Your audio is processed on your device and is not uploaded.

Privacy

Help

Keyboard shortcuts

Space / K
Play / pause
← / →
Back / forward 5 seconds
Shift + ← / →
Back / forward 10 seconds
Home / 0 · End
Start · end of song
M
Mute / unmute all audio
? · Esc
Open · close help

Playback shortcuts work when audio is ready. Space activates a focused button; arrow keys adjust a focused volume slider. Shortcuts pause while typing or viewing help.

Choose an audio file, then start processing. Your audio stays on your device.

Use Solo to hear a track on its own, Mute to silence it, and the volume slider to adjust the mix. Click a waveform or drag the seek bar to move through the song.

Chord Finder wraps the song into continuous timeline rows. Click a chord to jump to it. Detected chords are estimates and may need correction by ear.

Audio format support depends on your browser. Vocal removal and stem splitting currently require WebGPU. Keep this tab open while processing.

Models are downloaded once and reused while your browser retains its cache. Privacy