Vocal Remover
Take the singer out for a karaoke instrumental, or pull the vocal out on its own for an a cappella. AI stem separation splits the voice from the rest of the track: no phase tricks, no gutted bassline, no install.
Mute the singer, or keep only the singer.
AI separation puts the voice on its own channel, so an instrumental is one mute away and an a cappella is one solo away. No phase tricks, no scooped bass.
Upload a song & remove vocalsStep by step
- Upload the song
MP3, WAV, M4A or OGG. A high-bitrate source beats a stream rip.
- Let it separate the stems
Vocals, drums, bass, guitar, piano and other, usually in under two minutes.
- Mute or solo the vocal
Mute for karaoke, solo for the a cappella, then export lossless WAV.
How AI vocal removal actually works
For years, "vocal remover" meant a phase-inversion trick: flip one stereo channel, sum it with the other, and anything panned dead-centre cancels out. Because the lead vocal usually sits in the centre, some of it disappeared, but so did the kick drum, the bass, and any centred synth. You got a hollow, mono-ish instrumental with the low end scooped out. It was a compromise nobody was happy with.
This tool does not touch stereo phase. It separates your song into individual stems using an AI model trained on thousands of tracks with known parts. The model has learned what a human voice looks like in a spectrogram (its formants, its vibrato, the way it moves against a backing) and it isolates the vocal as its own track. Removing it is then simply a matter of muting that one channel, which leaves the drums, bass, guitar and piano completely intact.
The same separation gives you the reverse for free. Instead of muting the voice, solo it, and everything else drops away to leave the raw a cappella, useful for remixing, for studying a singer's phrasing, or for lining your own take up against the original.
What you can use it for
- Karaoke and covers: sing or play over a real instrumental instead of a MIDI backing.
- A cappella stems: isolate the voice for a remix, mashup or sample.
- Practice and analysis: solo the vocal to study phrasing, timing and ornamentation up close.
- Rehearsal tracks: mute the vocal so a live singer has a clean bed to perform over.
Related tools
Frequently asked questions
Does this leave the drums and bass intact when I remove the vocal?
Yes. The vocal is separated as its own stem, so muting it leaves every other instrument untouched, unlike old phase-cancellation removers that also gutted the kick and bass sitting in the centre of the mix.
Can I get the a cappella as well as the instrumental?
Yes. Mute the vocal for a karaoke instrumental, or solo it to hear it on its own. Both come from the same separation pass, so you can export either or both as WAV.
Why does a loud modern pop track leave a faint vocal ghost?
Brick-walled, loudness-maximised masters leave almost no headroom between the vocal and the mix, so there is less room to separate them cleanly and a faint "watery" residue can remain. Dynamically open recordings separate far more cleanly.
Is this better than a phase-inversion vocal remover?
Substantially. Phase inversion only cancels centre-panned sound and takes the bass and kick with it. This uses a trained model that recognises the vocal by its spectral signature, so it removes the voice specifically and keeps the rest of the arrangement.
Ready to lose the vocal?
Upload any track and get a clean instrumental or a cappella in minutes.
Open JamScore AI