MP3 Compressor
Compress MP3 audio to a size you choose. Lowers the bitrate to hit your target, dropping to mono and a narrower band only if the target needs it.
Up to 50 MB per file. Sign up free to compress files up to 100 MB.
| File name | Size | Compress to |
|---|
Checking files and selected compression settings.
This page stops at 50 MB and one file per run. Pro compresses files up to 1 GB and a full batch in one go, with a priority queue.
How to compress MP3 files online
MP3 is already lossy, so compressing one means re-encoding audio that has been encoded once before. The result depends far more on the bitrate you choose than on anything else.
-
1
Check the bitrate first
A podcast recorded at 64 kbps mono is already small and has nothing left to give. A 320 kbps album track has a great deal. Knowing which you have decides whether this page helps you at all.
-
2
Pick a target size
Size and duration are linked: a 10 MB target is generous for a three-minute song and severe for a two-hour lecture. The encoder lowers the bitrate until the file fits, so a long recording lands at a lower quality than a short one.
-
3
Listen before you keep it
Play the busiest passage rather than the opening. Cymbals, applause and room reverb are where a second encode shows first, as a thin, swirling quality around the high frequencies.
What MP3 compression actually costs
Every MP3 encode discards detail permanently. Re-encoding an MP3 starts from audio that is already missing information, and the second pass cannot recover it. Where the original WAV or FLAC still exists, encode from that instead.
A voice recording holds up at bitrates that would ruin a piece of music. Speech occupies a narrow frequency range with no stereo detail to protect, which is why interviews and lectures compress so much further than songs.
A lecture, a phone call or a dictation recorded in stereo is carrying two nearly identical channels. Converting to mono removes one of them before any quality reduction is needed, and nobody hears the difference on a voice recording.
Title, artist, album and artwork are preserved, so a compressed library stays organised. Embedded cover art can itself be a meaningful share of a short track's size, which is worth knowing when a file refuses to shrink.
File size is bitrate multiplied by duration. You cannot put an hour of audio into a few megabytes without dropping to a bitrate that makes music unpleasant, so for long recordings the honest options are mono, a lower target, or splitting the file.
If the file is going to a service that re-encodes on upload, as most podcast hosts and messaging apps do, your compression is discarded and replaced by theirs. Send the best version you have and let them do it once.