AUD · Audio tools

Extract Audio from Video

Shareable link

Settings are written to the URL as you change them. Nothing differs from the defaults yet.

Take the audio out of a video

Load an MP4, MOV, MKV, WebM, or MPEG-TS file and write its audio track to an audio file. The video track is discarded and the waveform of the audio is drawn so you can hear what you are about to save.

Two paths lead out of here, and the difference matters.

Copy vs re-encode

Original copies the encoded audio packets straight into a new container. Nothing is decoded, so nothing is lost and the work finishes at read speed. An MP4 holding 256 kbps AAC produces an .m4a holding that same 256 kbps AAC, byte for byte.

Everything else decodes the audio and encodes it again: WAV and FLAC without loss from that point, MP3, M4A, and Ogg Opus with a second lossy pass. Pick one of these when the destination cannot read the source codec, or when you need a specific bitrate.

The rule worth remembering: extracting to MP3 from a video is two lossy generations, not one. If the file is going into an editor, take Original or WAV.

When Original is not offered

The option appears only when the source codec fits a container this tool can write. AAC lands in .m4a, MP3 in .mp3, Opus and Vorbis in .ogg, FLAC in .flac, and PCM in .wav. Screen recordings, phone video, and most downloads fall into one of those.

MKV rips are where it stops. AC-3, E-AC-3, and DTS are common in them and none can be muxed here, so those files get the decode path and the format picker.

Codec and container are separate things

.mp4 names a container, not a codec. What sits inside it is usually AAC, which is why the copy path so often writes .m4a: same codec, different wrapper. The Stats row under the controls reports the codec the probe found, so the choice is made with the file in front of you rather than from the extension.

WebM works the same way. It carries Opus or Vorbis, both of which belong in an Ogg container, which is what the copy writes.

Take a section

Drag across the waveform to mark a range, tick the range option, and only that span is written. Useful for lifting a quote out of an interview or a music cue out of a longer recording without exporting the whole thing first and trimming afterwards.

Silent video

A file with no audio track has nothing to extract, and the probe says so instead of writing an empty file. Some screen recordings are captured without audio, and some MKV rips keep their audio in a track the container does not mark as primary.

Frequently Asked Questions

MP4 and M4V, QuickTime MOV, Matroska MKV, WebM, and MPEG-TS. AVI, WMV, and FLV have no demuxer here, so they are not accepted rather than failing after the upload.

It copies the encoded audio packets out of the video container and writes them into an audio container, with no decode and no re-encode. The samples in the output are the samples the video already held.

The audio codec has to be one an audio container here can hold. AAC, MP3, Opus, Vorbis, FLAC, and PCM all work. AC-3 and DTS, common in MKV rips, do not, so those files can only be decoded and re-encoded.

Yes. The video's audio is already lossy, and encoding it again to MP3 is a second lossy pass. Choose Original where it is offered, or WAV or FLAC if the file is going into an editor.

Drag a range across the waveform, then tick the range checkbox before extracting. The copy path trims on the container's own timestamps.

The video is read in pieces rather than loaded whole, so length is not the constraint. Output is written to origin-private file storage where the browser offers it, which is what keeps a long film from filling a tab's memory.

It is discarded. The output is an audio file, and nothing here writes video.

Know the format you started with?

Jump straight to an extraction

Explore Our Tools

Browse all tools