Where M4A files come from
Almost everyone who needs to convert M4A to MP3 got the file from one of three places. Voice Memos on an iPhone, which saves recordings as M4A, and the iPhone voice memo to MP3 route is by far the busiest one. iTunes and the Music app, which have used M4A for unprotected purchases and imports for years. Or a phone or camera app that writes audio-only MP4 without telling you.
Nothing is wrong with the format. It is simply less universally accepted than MP3, and the gap shows up in the annoying places: car head units, gym equipment, transcription services with a fixed upload list, DAWs on older versions, and any piece of hardware from before roughly 2010.
This conversion goes downhill, slightly
It is worth being clear about the direction. AAC in an M4A is a lossy encode, and MP3 is a different lossy encode. Going from one to the other means decoding what survived the first pass and letting a second encoder make its own decisions about what to discard. AAC is the better codec of the two at any given bitrate, so you are trading a small amount of fidelity for compatibility.
The practical advice follows from that. For music, use 256 or 320 kbps and the result will be indistinguishable in normal listening. For speech, 128 kbps is already generous and 64 kbps is defensible, because a voice has far less information in it than a full mix.
What the tool preserves
The sample rate is left alone. A 44.1 kHz source produces a 44.1 kHz MP3, and a mono recording stays mono rather than being padded into stereo. Both of those are easy to get wrong, and both matter for transcription tools that assume the file tells the truth about itself.
Everything runs on your machine. The engine is ffmpeg compiled to WebAssembly, downloaded once and then served from your browser cache.