Separate any audio into vocals and accompaniment using the Spleeter AI model. Upload a song or audio file — we split it into two tracks: clean vocals and instrumental. Everything runs locally in your browser for maximum privacy.
No. All processing happens locally in your browser using WebAssembly. Your files never leave your device.
Spleeter 2-stems separates audio into two tracks: vocals (singing, speech) and accompaniment (instruments, music). This is ideal for creating karaoke tracks or isolating vocals.
The model needs to process the entire file in memory. Files up to ~10 minutes work well on most devices. Longer files may be slower or require more RAM.
We use Spleeter 2-stems (fp16), originally from Deezer, running via sherpa-onnx WASM. It is one of the most popular source separation models.
Yes! Upload a video and we automatically extract the audio, separate it, and provide two downloadable tracks.