Remove the sound from a video

This tool writes a new file containing only the video track of the one you choose. The compressed frames are copied across without being decoded, so the picture is bit-for-bit what it was. The audio is not turned down or silenced — it is simply not in the output, which is the difference that matters if the reason you are here is that someone can be heard in the background.

Muting and removing are not the same thing

Setting the volume to zero in a player, or ticking a mute box in an editor that then exports the same audio at zero gain, leaves the recording inside the file. Anyone who opens it in a different program can turn it back up. Editors sometimes go further and re-encode a track of silence, which is worse than useless: you pay for a full re-encode and the file still has an audio stream.

Removing the track means the file has no audio stream at all. There is nothing to unmute and nothing to recover. If you are publishing a screen recording that caught a phone call, sharing a clip with a bystander conversation in it, or preparing footage that will get its own soundtrack later, that is the operation you actually want.

You can verify it. Any media inspector — the details pane in your file manager, VLC, ffprobe — will list the streams in the result and there will be exactly one. The page also tells you how many packets it copied, and for a muted output the audio count is zero.

Why the picture does not degrade

The usual way to remove audio is to load the file into an editor and export it. Most editors decode the video and compress it again on the way out, even when you changed nothing about the picture. That is a full transcode: it takes minutes on a long clip, it makes the file bigger or blurrier depending on the settings, and the loss is permanent.

None of that is necessary. A container stores each track separately, so the audio can be left out while the video frames are copied straight through in their existing compressed form. There is no decoder and no encoder in the path. A 6-second 720p clip takes about 5 ms here, which is the fastest of all the operations on this site, because it is doing the least work of any of them.

The output size drops by roughly the size of the audio, which is usually a small fraction of the file. If you were hoping removing the sound would make a large video much smaller, it will not — audio is typically a few per cent of the bytes. Making the video itself smaller is a re-encode, which we have not built.

What to expect from the result

The container stays in the MP4 family, which is what platforms and editors want. Rotation metadata is preserved, so a video shot on a phone in portrait does not come back on its side. Duration and frame timing are unchanged, so if you are going to line the clip up against a separate audio track, the sync points are where you left them.

If your source has more than one audio track, all of them are gone in a muted output, which is the intent. If it has more than one video track — unusual, but it happens with some screen recorders and camera apps — only the first is carried across, and the page says so rather than dropping the rest quietly.

A word on playing the result. If your device has no decoder for the codec inside the video, nothing on that machine will show the picture, including this page. That is a limitation of your machine, not a sign that the file is broken: copying compressed frames between containers never needs a decoder, which is also why this works on HEVC files from an iPhone even on hardware that cannot play them. The page checks before you start and tells you, instead of showing you a black rectangle.

Privacy, and why local processing is the point here

The reason people search for this tool is often the audio itself: a name, an address, a conversation that should not be in a file about to be published. Uploading that recording to a stranger's server to have it removed is a strange way to solve the problem. The audio you are trying to get rid of is the part you would be sending.

Here the file never leaves the tab. There is no upload request to inspect, no retention window to trust, and no download link that could be shared. Open your developer tools, watch the Network tab, and run the operation: your video appears in no request. The Content-Security-Policy header this site sends restricts connections to our own origin, so the browser itself would block an attempt to send it.

There is no size limit, no queue and no watermark, because processing costs us nothing. The one real constraint is device memory: the output is assembled in the tab, so allow for the output size plus about a hundred megabytes of working space, roughly 500 MB on a desktop and half that on a phone. The page warns you if your file is above that instead of letting the tab fail silently.