Choose a tool
Pick voice removal or a video workflow.
Remove vocals from any song to get a clean instrumental or acapella, or remove background music from a video while keeping the voice. AI separation in your browser — preview every track before you download.
Start with the sound you want to change. Switch tools in your workspace whenever your project calls for it.
Create a karaoke backing track or isolate vocals for a remix.
Open tool VIDEO MUSIC REMOVERSeparate the music bed from dialogue and surrounding sounds.
Open tool KEEP VOICE · REMOVE MUSICPreserve speech and effects while removing background music.
Open toolGet into the editor, choose a tool, and preview the separated tracks before downloading what you need.
Go to your workspacePick voice removal or a video workflow.
Upload a supported audio or video file.
Listen to the tracks and download your result.
Need help with your audio workflow? Start with these answers.
Songs are split into two tracks: the isolated vocal and the instrumental. Videos are split into three: dialogue, background music, and other sounds. For video you choose which layers to keep and get a finished MP4 with the original picture.
Songs are processed with htdemucs_ft, the fine-tuned Hybrid Transformer Demucs model for music source separation. Video soundtracks are processed with DBS-TIGER, a dialogue, music, and effects separation model trained on the Divide and Remaster dataset.
Audio: MP3, WAV, FLAC, M4A, AAC, and OGG up to 80 MB. Video: MP4, MOV, WEBM, and M4V up to 90 MB.
No. The video stream is copied into the finished file without re-encoding; only the audio track is replaced.
No. Processing runs on cloud GPUs and works in any modern browser. Sign in to process files and keep your results in a private file list.
Open your workspace and choose the audio tool that fits your project.