Add your file
Upload an MP4, MOV, WEBM, or M4V video up to 90 MB.
Choose which audio layers stay in your finished video.
For interviews, lessons, vlogs, and event footage: keep every word that was said — and, if you want, the room sound and effects — while the music underneath disappears.
Upload a fileUpload an MP4, MOV, WEBM, or M4V video up to 90 MB.
The soundtrack is split into dialogue, background music, and other sounds.
Listen to each layer and download the finished video with the original picture.
The soundtrack is split into three layers: dialogue, background music, and other sounds such as footsteps, applause, traffic, or a door closing. Most people keep dialogue and other sounds so the video still feels natural. Keep dialogue only when you want a clean voice track for voice-over work or transcription.
Your video is uploaded to private cloud storage and processed on a GPU worker. The soundtrack is separated with a dialogue, music, and effects model, the layers you selected are mixed together, and that mix replaces the original audio in a copy of your video. The picture stream is not re-encoded.
You also get all three layers as WAV files, so you can rebalance them in your own editor if the automatic mix isn’t quite right.
No. The video stream is copied into the finished file without re-encoding, so the picture keeps its original resolution and quality. Only the audio track is replaced.
Three separate WAV tracks — dialogue, background music, and other sounds — plus a finished MP4 that combines your original picture with the tracks you chose to keep. The finished video’s audio is encoded as AAC at 192 kbps.
MP4, MOV, WEBM, or M4V files up to 90 MB. For longer recordings, trim the part you need first or export a smaller file from your editor.
The separation model works on a mono version of the soundtrack, so the finished audio is centered rather than stereo. For dialogue-led videos this is rarely noticeable; for music-heavy edits with wide stereo effects, check the result before publishing.
Separating a track does not change who owns it. Removing vocals for personal practice is different from publishing the result. Only upload files you have the right to process, and get permission before you publish or monetize anything based on someone else’s recording.