Extract Audio from Video

Getting the track out

1. Put the clip in the Video area at the top. MP4, MOV, WebM and MKV all read fine, and the picture is ignored from that point on.
2. Read the strip: it names the sound track the video carries and whether the chosen format copies it or encodes it again. The line under the formats says which one copies.
3. Choose one of the four formats. M4A copies an AAC track, Opus copies an Opus one, and the choice is remembered for next time.
4. Press Extract, watch the run row, then play the audio take in the same window before saving it. The file keeps the video's name with the new extension.

About this tool

The fastest way to extract audio from a video is to not re-encode it at all. Feed in an MP4 or MOV whose soundtrack is AAC, leave the format on M4A, and the audio packets are lifted out of the video container and dropped into an audio one exactly as they were. Nothing is decoded and nothing is compressed a second time, which is why a ten-minute clip finishes almost as fast as you can click.

Ask for WAV, FLAC, or Opus and it does have to decode and re-encode, because those are different codecs from whatever came in. That is the price of a lossless editing file, or of a smaller one. The engine underneath is WebCodecs, the same hardware-backed decoder your device uses for its own camera, so even that slower path moves quickly.

People mostly reach for it to save the song off a clip, to keep the voiceover from a screen recording, or to get an interview they filmed into an audio editor. For the opposite job, Mute Video keeps the picture and throws the sound away.

A gigabyte of interview, kept as twelve megabytes of voice

A twelve-minute interview filmed on a phone arrives as an MP4 pushing a gigabyte, nearly all of it picture. Its soundtrack is AAC at 128 kbps, so with the format left on M4A the extract is a packet copy: the page says Copied before you press anything, the bar crosses in seconds and interview.m4a lands at about 12 MB, the voice untouched because nothing was re-encoded.

Headed into an editor instead? Press WAV and the same track is decoded to raw samples on the way out, well over ten times the size on disk, and the right foundation for cutting and levelling that ends in its own export.

Good to know

• There is no track picker. A file carrying more than one audio track, say a second-language dub, gives you no way to choose between them.
• MP3 is missing from the output list. Encoding it needs a licensed encoder that is not bundled here, so the choices are M4A, Opus, WAV and FLAC.
• A silent video produces nothing, and the page says so the moment it loads: the formats and the Extract button stay off rather than handing over an empty file.
• No trimming. The soundtrack comes out at full length, so if you only want thirty seconds of it, that is a second step in the Audio Cutter.
• Large files lean hard on memory. The output is assembled in a buffer before it reaches you, so a two-hour recording that a laptop swallows without noticing can run a phone out of room.
• Files only, never links. The page reads a clip already on your device and does not fetch from a URL, and what you may save off a streaming site is between you and that site's terms.