User Guide
Add the MP4
Drop the file in or click to browse. This tool also accepts audio files, so an M4A or WAV can go in the same way. The limit is 300 MB.
Wait for the engine to load
A status line reports when FFmpeg is ready — about 30 MB on the first run only. After that it starts immediately, including for later files in the same tab.
Press Convert
There is nothing to configure. The video track is discarded and the audio is encoded to MP3 at LAME quality V2, which averages around 200 kbps on typical material.
Expect it to be quick
Audio extraction is far faster than video conversion because no picture is ever decoded. An hour-long recording usually finishes in a couple of minutes.
Download the MP3
The file plays in every music app, car stereo, podcast player and DAW in existence — that universality is the whole reason MP3 is still the sensible default.
If you need a different format
Use the Video to Audio Converter for WAV, AAC or OGG output, or the Audio Cutter if you only want part of the recording.
About the MP4 to MP3 Converter
This tool takes the soundtrack out of an MP4 and saves it as an MP3, without uploading anything. It is the right tool when you want the audio from a recording and you want it to play absolutely everywhere.
What extraction actually involves
An MP4 is a container holding separate video and audio streams. Pulling the audio out means discarding the video track (the -vn flag) and dealing with what remains. Inside an MP4 that audio is almost always AAC.
Here is the part worth understanding: this is a transcode, not a copy. AAC is lossy and MP3 is lossy, and they discard different information. Converting one to the other means a second generation of loss on audio that already went through one. At V2 quality the difference is inaudible on normal listening — speech, podcasts, lectures, music through ordinary speakers — but it is not free, and it is the reason to keep your original if the recording matters.
If you want to avoid that second generation entirely, choose AAC output in the Video to Audio Converter. You will get the same codec family the MP4 already used.
The quality setting, and why it is fixed
The tool encodes at LAME variable bitrate quality 2. VBR spends bits where the audio is complex and saves them where it is simple, which produces better sound per megabyte than a fixed bitrate. The LAME encoder documentation states that on typical music V2 averages around 200 kbps.
| LAME setting | Average bitrate | In practice |
|---|---|---|
| V0 | ~245 kbps | Archival; larger files for little audible gain |
| V2 (used here) | ~200 kbps | Widely treated as transparent for most listeners |
| V5 | ~132 kbps | Acceptable for speech, audible on music |
V2 is not offered as a choice because it is the answer for essentially every use this page gets. Going higher enlarges the file without an audible improvement; going lower saves space you do not need on a format this efficient.
How big the result will be
At roughly 200 kbps the arithmetic is simple: MB ≈ minutes × 1.5. A 40 minute lecture is about 60 MB. A 3 minute song is about 4.5 MB. The size of the source video is irrelevant — a 250 MB MP4 and a 25 MB MP4 of the same length produce nearly identical MP3s, because the video track (almost all of the original file) is thrown away.
What people use this for
Recorded lectures and webinars that are easier to revise from on a phone. Conference talks and interviews for offline listening. Music videos where only the track is wanted. Screen recordings where the narration needs to be edited separately. Sermon and meeting recordings for archiving at a fraction of the size.
One practical note on that last group: if you only need part of a long recording, cut it first with the Audio Cutter. Extracting an hour and then discarding 50 minutes wastes both time and disk.
Privacy and limits
The extraction runs through a WebAssembly build of FFmpeg loaded into the browser tab on first use (about 30 MB, cached afterwards). Nothing is transmitted — there is no upload step and no copy of your file on any server. Because a single-threaded WebAssembly module addresses only about 2 GB of memory, files over 300 MB are rejected and anything above 60 MB warns that it will be slow. Since the video track is discarded rather than decoded, extraction runs much faster than any video conversion.
Frequently Asked Questions
What quality will the MP3 be?
LAME variable bitrate quality 2, which averages around 200 kbps on typical material according to LAME’s own documentation. That is widely treated as transparent — most listeners cannot distinguish it from the source on normal equipment.
Does extracting audio lose quality?
Slightly. MP4 audio is normally AAC, and both AAC and MP3 are lossy formats that discard different information, so converting between them costs a second generation. At V2 it is inaudible in ordinary listening, but choose AAC output in the Video to Audio Converter if you want to avoid it entirely.
Why can I not choose the bitrate?
Because V2 is the right answer for nearly every use of this page. Higher settings enlarge the file without an audible improvement; lower ones save space you do not need. If you want a different format or a smaller file, the Video to Audio Converter offers the choice.
How large will my MP3 be?
About 1.5 MB per minute — so a 40 minute lecture is roughly 60 MB. The size of the source video barely matters, because the video track is discarded rather than converted.
Can I put an audio file in instead of a video?
Yes. This tool accepts audio files as well as video, so an M4A, WAV or similar can be converted to MP3 the same way. For legacy formats such as WMA, FLAC or AIFF, use the Audio Converter.
Why is this faster than converting a video?
Because no picture is ever decoded. The video track is discarded outright, so the work is limited to the audio stream — a fraction of the data. An hour-long recording usually finishes in a couple of minutes.
Does my MP4 get uploaded for processing?
No. FFmpeg runs as WebAssembly inside your browser tab. The file is read from your disk into memory and the MP3 is written straight back out, so nothing is transmitted or stored.
Can I extract only part of the recording?
Not with this tool — it converts the whole file. Use the Audio Cutter to set a start and end point, which is also faster than extracting an hour and then trimming it.