--- title: "Remove Music from Video and Keep the Voice | ToolSugar" description: "Strip background music from a video and keep the speech. AI separation runs on your device, the picture is not re-encoded, and nothing is uploaded." canonical: https://toolsugar.com/tools/remove-music-from-video image: https://toolsugar.com/og.png --- # Remove Music from Video ToolSugar removes background music from a video with an MDX-Net separation model that runs in your browser, so the file is never uploaded. You get the video with only the voice, plus separate voice and music MP3s. The picture is copied without re-encoding. Files can run up to 15 minutes, and the first run downloads a 66.8 MB model. Runs on your device ## How it works 1. Choose a video file up to 15 minutes long. MP4, MOV, WebM, MKV and AVI all work. 2. Press Remove music. The first run downloads the 66.8 MB model, and later runs reuse the saved copy. 3. Wait while the audio is separated section by section. Keep the tab open until it finishes. 4. Download the video without music, or take the voice and music as separate MP3 files. ## Key facts - **Model**: UVR-MDX-NET Voc_FT, an MDX-Net model from the Ultimate Vocal Remover project - **First download**: 66.8 MB, saved in your browser for later runs - **Length limit**: 15 minutes per file - **You get**: The video, a voice MP3 and a music MP3 - **Picture**: H.264 and HEVC copied bit for bit with no quality loss, VP8 and VP9 re-encoded to H.264 - **Audio in the video**: MP4 with AAC 192 kbps, stereo at 44.1 kHz - **Speed**: On an Apple M4 Pro: about 16 s per minute of audio with WebGPU, about 3 min 45 s without it ## Common uses - Clean up a talk or interview filmed in a venue with music playing. - Take the backing track off a voiceover so you can add a different one. - Prepare a clip for subtitles or translation, where music makes the speech harder to follow. - Reuse an old presentation whose licensed music you no longer want in it. ## Questions ### Is my video uploaded anywhere? No. The audio is decoded and separated inside your browser tab. The only network traffic is the download of the model and the processing engine from public CDNs. ### Are sound effects removed too? Yes. The model sorts audio into voice and everything else, so door slams, traffic and applause go with the music. Sound effects you want to keep will be in the music file, not in the video. ### What happens to singing in the video? Singing counts as voice, so a singer stays in the output along with anyone speaking. This tool cannot keep the speech and remove only a song's vocals. ### Will the result be perfectly clean? Not always. Heavy reverb, live room recordings and music much louder than the voice can leave a faint trace of music, or make the voice sound thinner. Listen to the result before you publish it. ### Does removing the music solve a copyright claim? Not by itself. Automated systems may still match what is left, and the rights in the rest of the video are unchanged. If a claim matters, check with the platform or the rights holder. ### How long does it take? We measured it on an Apple M4 Pro. With WebGPU, a minute of audio took about 16 seconds and a 15-minute video took about 4.5 minutes. Without WebGPU, a minute of audio took about 3 min 45 s, so a 15-minute file would take close to an hour. Slower computers take longer, and we have not tested phones. ### Which model does this use? UVR-MDX-NET Voc_FT, an open vocal separation model. It runs in your browser, so the video stays on your device. [Read the Remove Music from Video guide](https://toolsugar.com/blog/how-to-remove-music-from-a-video) ## Related tools - [Extract Voice from Video](https://toolsugar.com/tools/extract-voice-from-video) - [Remove Music from Audio](https://toolsugar.com/tools/remove-music-from-audio) - [Vocal Remover](https://toolsugar.com/tools/vocal-remover) - [Remove Audio](https://toolsugar.com/tools/remove-audio-from-video) - [Auto Subtitles](https://toolsugar.com/tools/subtitle-generator) - [Noise Remover](https://toolsugar.com/tools/noise-remover)