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: August 2026
Music mixed into your narration is a copyright problem waiting to happen, and muting it in your editor mutes your voice too. Drop the video in below: the same AI that pulls a clean acapella out of a song lifts your narration out of the music behind it, and hands you the two apart.
One upload, two tracks:
Drop the voice track into your video editor in place of the original audio and the background music is gone, without re-recording a thing.
Either works. This tool takes video files (MP4, MOV, MKV, WEBM) directly, so for most people the fastest path is dragging the video straight in above. If you'd rather work from an audio file — say you already pulled the sound out for something else — use extract audio from video first to get an MP3, then run that MP3 through this same tool.
Most video edits — a vlog with a licensed track under it, a tutorial with intro music, a voiceover laid over stock footage — end up flattened to a single stereo audio track when the project is exported. If you still have the original project file with the voice and music on separate layers, muting the music layer is trivial. If you only have the finished video, or you're re-using a clip someone sent you, the two are stuck together and this is the tool that pulls them back apart.
The model separates music from voice, the same as it does on a song. It doesn't remove non-music background noise like wind, room hum or a fan — for that, run the voice track it gives you through a noise reducer like CleanGrab afterwards. Dense, loud music under a quiet voice is harder than a clear voiceover over a subtle bed track, and heavy reverb on the voice can leave a faint trace behind, same as with music separation on a song.
No — drop the video file straight into the tool above and it separates your voice from the music in one pass. If you'd rather work from an audio file, use extract audio from video first, then run the MP3 through this tool.
No, this separates music from voice, not noise from voice. For hum, wind, hiss or room echo, use a noise reducer like CleanGrab instead — you can run that afterwards on the voice track this tool gives you.
Because in most videos the music is mixed into the same track as your voice, not on a separate layer — muting it mutes you too. This lifts the two apart after the fact, which is exactly what you need when the original project file with separate tracks is gone.
Yes — no account, no e-mail, no daily cap and no watermark. The only limit is a file up to about 100 MB and 15 minutes long per pass.
It's run through the same model that isolates vocals from songs, so it stays close to your original recording. Dense background music, heavy reverb or a lot of overlapping talking in the clip make it harder for the model, same as with a song.
No. Your upload is deleted the moment it's processed and the stems are wiped from the server within 45 minutes. Nothing is stored, shared or used to train anything — see privacy.