This free karaoke maker turns any song into a clean instrumental in about six seconds for a three-minute track — the lead vocal lifted out, not just filtered away. No account, no watermark, free in the browser.
Full-quality stems. No account, no per-track fee. Same model and same 192 kbps (or lossless WAV) as the paid tier — tracks up to 15 minutes.
By ShinobiTools Team · Last updated: July 2026
This is a karaoke maker app that runs the same in a phone browser as on a desktop — a karaoke maker split finishes in about six seconds for a three-minute song. It's a karaoke maker free online tool on every device, with the same quality on free and paid.
Split your track free →Drop the song in — MP3, WAV, FLAC, M4A, AAC or OGG all work, up to 15 minutes and 2 GB per file, free. The AI splits it on our own GPUs, usually under a minute for a three-minute song; if it's busy you get a queue place, not an error. Download the instrumental — your karaoke track — plus the isolated vocal and five more stems, all as 192 kbps MP3 or lossless WAV.
Old karaoke tools cancel whatever sits dead-centre, taking the kick and bass down with the singer and only working on centred recordings. StemGrab doesn't cancel — it recognises what a voice sounds like, so mono and off-centre vocals are fine. A modern studio recording with the lead voice on top comes back cleanest; heavy harmonies, reverb, live crowd noise and loud masters all leave more vocal bleed — true for any separation tool. It also can't split a duet by singer: the AI lifts out one full lead-vocal track, not separate voices.
As a karaoke maker ai tool it recognises what a voice sounds like rather than cancelling the centre of the mix, which is also why it doubles as a karaoke maker vocal remover for the exact same upload.
This karaoke generator studies the mix to find the voice rather than cancelling the centre channel. As a karaoke track maker, the download includes the instrumental plus the isolated vocal and five more stems from one upload. Used as a karaoke creator, mono and off-centre vocals still separate cleanly. As a karaoke maker online free of any subscription, ten tracks a day come back at the same quality. To make karaoke online, the upload box above handles MP3, WAV, FLAC, M4A, AAC and OGG.
Run as a song karaoke maker, a three-minute upload usually finishes in about six seconds. As a karaoke converter, files up to 15 minutes and 2 GB go through free. To make karaoke from a duet, the tool lifts out one full lead-vocal track rather than separating each singer. To convert song to karaoke tracks with heavy reverb or a live crowd, expect a bit more vocal bleed.
As a karaoke creator online, the split runs on our own GPUs. This karaoke creator online free tier has no watermark and no account requirement. As an ai karaoke maker, the model recognises what a voice sounds like instead of cutting the stereo centre. A karaoke online free run gives back an MP3 by default, with lossless WAV on a paid pass.
As a karaoke song generator, harmony vocals and backing singers leave with the lead. To turn any song into karaoke form, a modern studio recording comes back the cleanest. To make a song karaoke track from a cover you recorded yourself, the upload works the same as any commercial release. Every karaoke online result here comes with the isolated vocal alongside the instrumental, not just the backing track alone.
To make karaoke from song uploads over 15 minutes, splitting the track first is the only workaround for the free tier. As ai karaoke, the tool doesn't add any effects or backing music of its own — only what was already in the original mix comes back. The short version of how to make song karaoke version is: upload it, wait a few seconds, then download the instrumental. For how to make a karaoke version of a song already in a set list, ten tracks a day covers a typical rehearsal session.
The karaoke ai split treats guitar and piano as newer, more experimental stems, which can come out rougher on dense mixes. As an ai karaoke generator, the download always includes seven files: the instrumental, the vocal, and five separate instruments. An mp3 to karaoke run keeps the original key and tempo of the upload, with nothing shifted or re-tuned. Used as a karaoke builder for a whole set list, each song still needs its own separate upload and download. For how to make karaoke tracks ahead of a party, batching a few songs through one at a time still takes about six seconds each. To create karaoke from an old recording, a lower-bitrate source file gives the model less to work with than a clean original.
Yes — no account, no watermark, 10 tracks a day. The ads around the tool keep the GPU running.
The tool is free to use, but the rights to the recording stay with whoever owns them. Practising privately is always fine; for anything you publish — a remix, cover or mashup — get permission first. See stems & the law.
StemGrab focuses on the separation itself, not editing. For trimming the finished track down to a shorter clip, run the download through the audio cutter — no extra upload of the original song needed.
Not on this page, but on the next one: the song key tool moves a whole track up or down in semitones, and the pitch changer does finer steps for when a song is only just out of reach. Make the instrumental here first, then shift that file into a key you can comfortably sing.
No. The upload is deleted right after it is processed, and the results are wiped from the server after 45 minutes, or 24 hours if you bought a pass. Nothing is stored, shared or used to train anything — see privacy.
Yes — StemGrab is a karaoke maker online with 10 tracks a day: upload any song and download the instrumental a few seconds later. No account, no subscription, no watermark.
No — backing and harmony vocals leave with the lead, since the model treats all singing as one part; what's left is the pure instrumental, not a lead-only removal.
For a normal studio mix, yes — the split finishes in about six seconds for a three-minute song, at the same quality on free and paid.
No editing needed — drop the song in and the instrumental downloads a few seconds later, ready to sing over.
Yes — upload the cover the same as any other song and the lead vocal is lifted out, leaving your own backing as the karaoke track.
Just one upload — drop the file in, wait a few seconds, and download the instrumental; there's no separate settings screen.
Yes — the separation doesn't care about genre; a modern studio recording with the lead voice on top comes back cleanest, whatever style it is.
MP3 works directly, and so do WAV, FLAC, M4A, AAC and OGG — no converting needed first.
Upload your own copy of the English-language recording and it comes back the same as any song — we don't host or provide the file ourselves, just the separation.
Only the browser — Chrome, Edge, Firefox or Safari on Windows or Mac all work the same, with nothing to install.
Not easier, just different — CapCut needs the app installed and a project opened; this page only needs the song file and gives back an instrumental MP3.