RunToolRunMP4 to MP3
Convert MP4 files to MP3 format.
How to Extract MP3 Audio from Video
Select a video file
Click the upload area or drop an MP4, MOV, WebM or other video.
Audio is extracted locally
The video track is discarded and the sound is encoded as high-quality MP3 in your browser.
Download the MP3
Save an audio file that's a fraction of the video's size.
Extracting audio is not converting video
An MP4 is a container holding separate video and audio streams alongside timing information. Extracting the audio means reading the audio stream out and writing it into an MP3 file — the video is simply left behind rather than transformed into sound.
This is why the operation is fast relative to video conversion and why the result is so much smaller. A 200 MB lecture recording becomes a few megabytes of audio, because the overwhelming majority of a video file is pictures.
The source sets the ceiling
The audio inside an MP4 is usually AAC, which is already compressed. Writing it as MP3 means decoding and re-encoding, so this is a second generation of lossy compression and it cannot improve on what the video contained.
The practical consequence is about expectations rather than settings. Requesting a 320 kbps MP3 from a video whose audio track was recorded at 96 kbps produces a large file containing 96 kbps of actual quality. Match the output bitrate roughly to the source: 128–192 kbps is right for speech and ordinary video audio, and higher numbers only add size.
What MP3 gives you and what it costs
MP3 plays on absolutely everything — car stereos, old players, every phone, every editing program. That universality is why it remains the default request despite newer formats compressing better. For spoken content it is entirely sufficient.
It carries no video, obviously, but it also carries none of the chapter markers, subtitles or multiple audio tracks a video file may have had. If a recording has a second language track or timed markers you rely on, extracting to MP3 discards them.
Practicalities and rights
Extraction runs through a WebAssembly build of FFmpeg in your browser, so lectures, interviews, meeting recordings and anything else confidential are processed on your own machine. It is single-threaded, so a long recording takes real time — a two-hour file is a genuine wait rather than an instant operation.
Worth stating plainly: extracting audio does not grant any right to redistribute it. Converting your own recordings, licensed material or content you have permission to use is fine; pulling the soundtrack from someone else's work and publishing it is not, whatever the tool makes technically easy.
Speech and music want different settings
Spoken content — lectures, interviews, podcasts, meeting recordings — carries almost all its information in a narrow frequency band, so it survives aggressive compression. Mono at 96 to 128 kbps is usually indistinguishable from the source and produces a file small enough to keep on a phone indefinitely.
Music needs more headroom. Stereo imaging, cymbals and reverb tails are exactly what low-bitrate encoders smear, so 192 kbps or above is worth the space. Either way, remember the ceiling set by the video's own audio track: no output setting can add detail the source never contained.