Keep only the sound — copied out of the video untouched, or decoded to WAV.
Your files never leave your browser — all processing is local.
Drop a video
MP4, MOV, WebM, MKV — audio is copied in the browser, never uploaded
A lossless copy does not improve the audio and does not turn AAC into a truly lossless format. WAV is just more widely usable, and larger. DRM-encrypted files can't be read. Title, date and GPS tags are not copied.
Drop a video and the tool lists every audio track inside it — language, channel count and codec — so a film with two dubs or a screen recording with separate mic and system audio lets you pick the one you want. You can extract the whole thing or just a range, set the same way as in the Video Trimmer. The file stays on your computer: it is read by random access, never uploaded, and the result is written straight back to your disk.
The default mode copies the audio packets exactly as they are into an audio-only container: AAC becomes .m4a, Opus and Vorbis become .ogg, MP3 stays .mp3, FLAC stays .flac and PCM becomes .wav. Not a single sample is re-encoded, so the sound is identical to what was in the video and the job takes about as long as reading the file. The second mode decodes the track to PCM and writes a WAV file. WAV is large — the tool shows the estimated size before you start — but every editor and transcription program opens it without a codec question.
Most people arrive here searching for "MP4 to MP3", so here is the honest answer: MP3 requires re-encoding — decoding the AAC and encoding it again — and this tool does not offer that yet. What it gives you instead is the AAC track itself as an M4A, which is what almost every MP4 already contains. M4A plays on every phone, computer and browser, and nothing is lost in the process. If you need a file for editing or for transcription software, choose WAV.
A few limits are worth stating plainly. AC-3 and E-AC-3 audio — common in MKV files and TV recordings — cannot be decoded by most browsers, so the tool tells you instead of producing a file that would not play. A video with no audio track says so instead of writing an empty file. DRM-protected files cannot be read at all. And "lossless extraction" means nothing is lost on the way out; it does not turn a lossy AAC or Opus track into lossless audio. Once you have the audio, Video Subtitles can turn the speech into a transcript.
Not yet. An MP4's audio is almost always AAC, and turning it into MP3 means decoding and re-encoding — slower, and it costs a little quality. This tool copies the AAC out as an M4A instead, which plays on every phone, computer and browser and sounds exactly like the original. If your video already carries an MP3 track, you get an .mp3 file as-is. For a file that any editor can open, choose WAV.
In the default mode, no — the compressed packets are copied bit for bit, so the M4A, OGG, MP3 or FLAC you get holds exactly the audio that was in the video. WAV mode decodes the same packets to uncompressed samples; that does not lose anything either, but it does not gain anything: a lossy AAC track does not become better by being written as WAV.
After the file opens, every audio track is listed with its language tag, channel count (mono, stereo, 5.1) and codec. Select the one you want before extracting; each run produces one file, so repeat for another track if you need it. MKV files and some MOV recordings are where this usually matters.
Browsers generally have no decoder for Dolby AC-3 / E-AC-3, which is common in MKV files and TV recordings. Copying it out would give you a file that your browser and most phones cannot play, and converting it to WAV needs a decoder the browser does not have. So the tool stops and tells you which track is affected rather than producing something broken. A desktop tool such as ffmpeg or VLC can convert those tracks.
M4A (or OGG, FLAC, MP3 — whichever your video already contains) when you want to listen: it is small, it is the original audio untouched, and it plays everywhere. WAV when the file is going into an editor, a DAW or transcription software: it is uncompressed, so it is many times larger — the tool shows the estimated size first — but nothing needs to decode it.
No. The video is read from your disk in your browser and the audio file is written back to your disk. No server sees the file, which is also why a multi-gigabyte video opens immediately and there is no queue.
Yes. Set an in and out point on the timeline — by dragging, by pressing set-in / set-out during playback, or by typing a timecode — and only that range is written. The cut lands on an audio packet boundary, so it can differ slightly from the exact point you chose.