Which fix fits your video
| Situation | Best fix | You keep | You lose |
|---|---|---|---|
| Music under speech: vlog, interview, tutorial | AI separation on this page | Your voice | The song |
| Music only, no speech: B-roll, timelapse | Mute and add new music | Picture | All original sound |
| Video already on YouTube with a claim | YouTube Studio's Erase Song editor | Most other audio | The claimed segment |
| You sing along with the song | Separation keeps the singing; it is still the song's melody and words | Your vocal | Backing music only |
| Song plays loud through a speaker in the room | Separation, expect some leftovers | Most of the voice | Most of the music |
Separation is the right tool when the voice matters. If nobody speaks in the clip, muting is simpler.
What separation does, and what it cannot promise
The AI model, Demucs htdemucs, is trained to tell voices from instruments. For a voice over background music it usually removes the music cleanly. It is less clean when the song itself has vocals, because the singer counts as a voice too, and when music is very loud next to quiet speech.
- Content ID compares audio fingerprints. Faint leftovers can still match, so listen to the result before you upload.
- If a song's vocals were playing, they may remain in the Vocals stem. Cut those parts with the free audio cutter.
- Removing music does not change who owns the video. Only do this on footage you have the right to use.
- For a video already claimed on YouTube, YouTube's own Erase Song tool in Studio can remove the claimed song and may release the claim.
Put the clean voice back on your video
| Editor | Steps |
|---|---|
| CapCut | Tap the clip, Volume to 0 (or Extract audio and delete it), Audio, Add, pick the Vocals MP3, align at start |
| iMovie | Select clip, Detach Audio, delete it, drag the Vocals MP3 under the clip at 0:00 |
| Premiere Pro / DaVinci Resolve | Unlink, delete the original audio, import the MP3 to A1 at 00:00:00:00 |
| Clipchamp | Detach audio, delete it, add the MP3 to the timeline at 0:00 |
The stems start at zero and have the same length as the soundtrack, so the voice lines up as long as it starts at the very beginning.
Hear a real separation
This is a real 4-stem result from the engine behind the tool. For a video, the 2-stem mode gives you the Vocals file and a Music file.
Frequently asked questions
How do I remove copyrighted music from a video but keep my voice?
Upload the video to the tool on this page with 2 stems. AI separates your voice from the music and you can download your video back with the voice as its soundtrack (built in your browser, no second upload), or take the voice as an MP3 and replace the audio in your editor. 1 full split a day is free, up to 10 minutes.
Will this stop a YouTube copyright claim?
Often it removes the music well enough, but it cannot guarantee that. Content ID can match faint leftovers, and a song's own vocals stay in the voice track. Listen carefully before uploading. For a video that is already claimed, YouTube's Erase Song tool in Studio is the official route.
Do I get a video file back?
Yes. After the split, click Download video without music and you get an MP4 with the same picture and the voice as its soundtrack. It is built in your browser, so the video is not uploaded again; on a browser that cannot do it, download the voice MP3 and add it at the start of the timeline in CapCut, iMovie or any editor. The voice and the music are also available as MP3 files.
Can I remove music from a TikTok or Instagram video?
Yes, if you have the file. Save your own video from the app or your camera roll, then upload the MP4 here. Up to 10 minutes and 150 MB per file.
Is my video kept?
No. Files are stored privately, used only to make your stems, deleted within 24 hours and never used for training.
Does removing the music make someone else's video mine to use?
No. It removes a song from the soundtrack; it does not give you rights to the footage. Use it on videos you recorded or have permission to use.