Which file to keep for your job
A voice remover always gives you both halves. What changes is which one you download.
| You want | Upload | Download |
|---|---|---|
| A karaoke or backing track | The song | Instrumental |
| The background music from a video, without the person talking | The MP4 or MOV | Instrumental |
| A voice-over without the music under it | The video or recording | Vocals |
| Your singing without the backing track you sang over | Your recording | Vocals |
| A video with the music removed but the speech kept | The MP4 or MOV | Download video without music |
| An acapella for a remix | The song | Vocals |
If you need the drums or bass on their own too, switch to 4 stems before you upload.
Hear a voice removed from a song
Here is a real result on Sunlight, a song we generated ourselves. The Instrumental is the song with the voice removed; the Vocals file is what was taken out.
Speech works the same way as singing: the model hears a human voice and moves it to the vocals file. Spoken words over a quiet music bed usually separate very cleanly, often cleaner than a lead singer in a dense pop mix.
Removing a voice from a video
- Upload the MP4 or MOV straight from your phone or computer. The audio track is what gets separated.
- To keep the speech and drop the music, use the Download video without music button: it rebuilds your video in the browser with only the voice.
- To keep the music and drop the speech, download the Instrumental MP3 and add it back in CapCut, iMovie or any editor.
- Check the start and end. Fades and sound effects can split unevenly between the two files.
The video itself is never re-uploaded for the export step. The new file is built on your device.
What a voice remover can and cannot do
| Situation | Result |
|---|---|
| Lead singer over a band | Good; backing vocals go with the lead |
| Narration or talking over background music | Very good |
| A voice memo with a TV or music in the background | Good when the voice is louder than the background |
| Two people talking over each other | Both stay in the vocals file; it cannot pick one person |
| Wind, traffic or crowd noise | Not what it is built for; use a noise reducer |
| Heavy reverb or echo on the voice | The echo usually leaves with the voice |
For noise rather than music, the voice cleaner tool is the better fit. For removing music from a whole video, there is a dedicated page linked below.
Private by default
Choose In your browser and the file never leaves your device: the same AI model runs locally, with no daily limit for files up to 8 minutes. In Fast mode, files go to private storage, are used only for your result and are deleted within 24 hours. They are never used for training.
Frequently asked questions
What is a voice remover?
A voice remover is a tool that separates a human voice from the rest of an audio file. TuneSplit uses the Demucs htdemucs AI model to split any song, video or recording into two files: the voice and everything else. You can keep either one.
How do I remove a voice from a song for free?
Upload the song to the tool above with 2 stems and download the Instrumental MP3. It is free with no sign-up: unlimited in your browser with no upload, or 1 fast server split a day, up to 10 minutes.
Can I remove a voice from a video?
Yes. Upload an MP4 or MOV. To keep the music, download the Instrumental MP3 and put it back under the picture in any editor. To keep the speech and drop the music, use Download video without music, which rebuilds the video in your browser.
Does it work on speech, or only singing?
Both. The model was trained to find human voices, so narration, interviews and voice-overs separate from background music just like a singer does. Talking over a quiet music bed is often one of the cleanest cases.
Can it remove one person's voice and keep another?
No. All voices go to the vocals file together, so it cannot separate two speakers or two singers from each other. It separates voices from music and instruments.
Can I use it on iPhone or Android?
Yes. Open the page in Safari or Chrome, tap the upload box and choose a file from Files, Photos or Voice Memos. Fast mode works best on phones; browser mode needs more memory and time.
Can I split a song without uploading it, for free and without limits?
Yes. Choose In your browser above the drop zone. The same Demucs htdemucs model runs on your own device, so the file is never uploaded and there is no daily limit, for 2 or 4 stems and songs up to 8 minutes. The first run downloads the 172 MB model once. It is slower than the server, usually a few minutes for a 3-minute song, and works best in a recent Chrome, Edge or Safari on a laptop or desktop.
Is my file kept or used for training?
No. In browser mode it never leaves your device. In Fast mode it is stored privately, used only to produce your result, and deleted within 24 hours. Files are never used to train any model.