What extracting the audio actually does
A video file is a container with separate tracks inside: one for the picture and one or more for sound. MP4, M4V and MOV usually hold H.264 or HEVC video with AAC audio. WebM holds VP8, VP9 or AV1 video with Opus or Vorbis audio. MKV, AVI, WMV, MPG, 3GP and TS files wrap other combinations. In every case the audio is a track of its own, and this converter reads that track and writes it to a new file. The picture is left behind, which is why the output is a small fraction of the video's size.
The converter accepts any video FFmpeg can read, which covers all of the containers above, and it also accepts plain audio files if you want to change their format.
MP3, M4A or WAV: which to choose
Three output formats are offered. The right one depends on what you will do with the audio afterwards.
- MP3: the safe choice for listening and sharing. It plays on everything, and you can pick 128, 192 or 320 kbps. The audio is decoded and re-encoded, so a little detail is lost. For speech at 128 kbps or above you will not hear the difference.
- M4A: the audio-only MPEG-4 container, holding AAC. When the video already carries AAC audio, which is the case for almost every MP4 and MOV from a phone, camera, Zoom or a screen recorder, the audio stream is copied into the M4A without re-encoding. That is instant and lossless: the result is bit for bit the same sound that was in the video. When the source audio is something else, such as Opus in a WebM, it is encoded to AAC instead.
- WAV: uncompressed 16-bit PCM. Choose it when the audio is going into Audacity, a DAW or a video editor, or into a speech tool that wants 16 kHz mono. You can set the sample rate to Same as source, 48 kHz, 44.1 kHz or 16 kHz, and fold stereo to mono. WAV does not add detail to a compressed track, but it stops any further loss while you edit.
Who pulls audio out of video, and what they pick
Video editors often need the dialogue track on its own, to clean it up in a separate program or to hand it to a sound editor. For them M4A is the quickest route when the camera recorded AAC, and WAV is the format every audio application opens without decoding.
Podcasters who record remote conversations on video, through Zoom or a phone camera, usually want M4A or WAV for the edit and MP3 only at the very end. Interviews filmed for a documentary or a research project are the same: keep the lossless copy for the archive and make an MP3 when someone just needs to listen.
Students and teachers extracting a lecture from a screen recording mostly want MP3 at 128 kbps, which is small enough to send and plenty for speech. Musicians lifting a live performance from a phone video should choose M4A if the phone recorded AAC, or MP3 at 320 kbps if they need MP3 specifically.
Your video never leaves your device
This tool is FFmpeg compiled to WebAssembly, running inside the browser tab. The video is not uploaded, and nothing is stored on our side. That is the point for footage you cannot share: client interviews, unreleased footage, meetings with confidential content. On first use the browser downloads the conversion engine, about 32 MB, and caches it, so later conversions also work offline.
One file at a time, up to about 2 GB on a desktop or laptop. Phones and tablets manage less because the browser has less memory. How long a conversion takes depends on your device, and the M4A stream copy is the fastest option because nothing is re-encoded.
The output keeps the original file name with the new extension. Press Download to save it. The result exists only in the open tab, so save it before you reload or close the page. There is no account to create and no watermark, and you can process as many videos as you like. Only local files work, so a YouTube, TikTok or other video link cannot be used here. Current Chrome, Edge, Firefox and Safari 16 or newer are supported.
If you actually need the words
When the reason for extracting the audio is to get it transcribed, you can go straight there. After the conversion, press Transcribe this file and the audio moves into the transcription box further down this page. That step runs on our servers, so the audio is uploaded for processing. You get a timestamped, editable transcript in 90 or more languages, with exports to TXT, DOCX, PDF, SRT, VTT and CSV. The SRT or VTT file lines up with the original video, so you can load it back into your editor as subtitles. You can try it without an account.