One converter for every video container
Video files come in many wrappers, and most online tools handle only a few of them. This converter accepts anything FFmpeg can read, which covers every common container:
- MP4 and M4V: phones, screen recorders, Zoom, online courses. Usually H.264 video with AAC audio.
- MOV: Apple's QuickTime container from iPhones, iPads and Macs. Often H.264 or ProRes video with AAC or uncompressed PCM audio.
- WebM: browser recordings and web video. The audio inside is Opus or Vorbis.
- MKV: a flexible container popular for archived video. The audio can be almost any codec.
- AVI and WMV: older Windows formats from camcorders, capture cards and early editing software.
- 3GP: videos from older feature phones and some Android recorder apps.
- MPG, MPEG and TS: DVD material, broadcast captures and some camcorders.
Why the same conversion works for all of them
A video file is a container holding separate streams: one for the picture, one or more for sound, sometimes subtitles. Converting to MP3 does not involve the picture at all. The converter opens the container, finds the audio stream, decodes it to plain samples and encodes those samples as MP3 with the LAME encoder.
That is why the container does not matter. Whether the audio arrived as AAC in an MP4, Opus in a WebM, PCM in a MOV or an older codec in an AVI, the decoded samples look the same to the MP3 encoder. FFmpeg carries decoders for all of them, so one tool covers the whole list above.
It also explains two things people notice. First, the MP3 is much smaller than the video, because the picture is gone. Second, the conversion is only as good as the audio stream inside the file. A video with a quiet or distorted soundtrack produces an MP3 with the same flaws.
Picking a bitrate for speech or music
You can choose 128, 192 or 320 kbps. The right choice depends on what is in the soundtrack rather than on the picture.
Speech. Talks, lectures, meetings, interviews, tutorials and vlogs are mostly voice. Voice does not need much bandwidth: 128 kbps is plenty, and the files stay around 1 MB per minute. If the file is going to be transcribed, 128 kbps is also enough for that.
Music and mixed content. Concerts, music videos, film clips and anything with a wide stereo mix benefit from 192 kbps or 320 kbps. 320 kbps is the ceiling for MP3 and is the setting to use when you want to lose as little as possible.
Unsure. Leave the default of 192 kbps. It is good enough for music to most ears and still much smaller than the video.
Remember that the audio in most videos is already compressed with a lossy codec. MP3 is lossy as well, so each conversion removes a little more. A higher bitrate keeps the second loss small. It cannot recover detail the original codec removed. The exception is video with PCM audio, such as some MOV files and camcorder footage, where the MP3 step is the first lossy stage.
Everything happens on your device
The converter is FFmpeg compiled to WebAssembly, running in your browser tab. The video is read from your disk, converted locally and never uploaded. RocketScribe does not receive it and keeps nothing. That matters for private footage, client work and anything under NDA, and it means a large video does not have to travel over a slow connection before anything happens.
On first use the browser fetches the engine once, about 32 MB, and caches it. After that the page works offline. It converts one file at a time, up to about 2 GB per file on a desktop browser; phones and tablets handle smaller files. The result keeps the original name with an .mp3 extension and lives only in the open tab, so press Download before leaving the page. Supported browsers are current Chrome, Edge, Firefox and Safari 16 or newer.
If you actually need the words
If the video is a lecture, an interview or a meeting, the MP3 may just be a step on the way to a transcript. After converting, press Transcribe this file to send the MP3 into the transcription box below. Transcription runs on RocketScribe's servers, so the audio is uploaded for that step. It comes back as an editable, timestamped transcript in 90+ languages that exports to TXT, DOCX, PDF, SRT, VTT or CSV. No account is needed to try it.