Pull the audio track out of an MP4 (or MOV, WebM, MKV) as an MP3 — choose the bitrate, trim a section, 100% in your browser.
Drop a video to pull out its audio as MP3 — locally, no upload
Conversion runs entirely in your browser with ffmpeg.wasm — the video never leaves your device. Larger videos need more memory: the WebAssembly build works in a single thread and a 200 MB file can take several minutes.
An MP4 is a container: it holds a video track and an audio track that were encoded separately, usually AAC for audio. Extracting MP3 is therefore not “converting video into sound” — it is copying the audio stream out and re-encoding it to MP3. That distinction matters because it means quality is capped by the source audio: a video whose audio was encoded at 96 kbps will not improve at 320 kbps, it will just get bigger. The practical rule: 128 kbps is fine for speech (podcasts, lectures, interviews), 192 kbps is the safe default for music, and 256–320 kbps only pays off when the source audio is genuinely high quality. Mono halves the file again and is usually better for voice recordings, because a centered voice has no stereo information to preserve. Trimming with -ss/-t before encoding also saves real time: ffmpeg seeks first, so a 30-second clip from a 2-hour video encodes in seconds rather than minutes. Everything here happens locally via ffmpeg.wasm, which means no upload queue and no privacy question — but it also means the work is done by your CPU, so long videos take longer than a server would.
The video track is discarded (-vn); the audio stream is re-encoded with libmp3lame at your chosen bitrate.
A 52-minute recorded webinar needs to become an audio file small enough to listen to on a phone.