Convert Opus to MP3

Why your voice notes are Opus files

Opus is the best speech codec that exists at low bitrates. A clear, intelligible voice note fits into 16–24 kbps — roughly 150 KB per minute — which is why WhatsApp, Telegram, Signal, Discord, Google Meet, Zoom’s web client and essentially every modern real-time system use it. It is also royalty-free, which removed the last reason not to.

The problem is playback outside those apps. Windows Media Player, older iPhones, most car stereos, plenty of transcription services and a lot of legal and medical intake systems will not open an .opus file at all. Converting to MP3 is almost always about handing the recording to something else.

What the file will be called

All five are accepted here. If your file is Vorbis rather than Opus — game audio, a Spotify cache, a Wikipedia clip — it still converts, but OGG to MP3 is the page written for that case.

Bitrate: why 128 kbps is enough and 320 is waste

This is the conversion where people most often pick the wrong number. The instinct is to choose 320 kbps "to be safe", which produces a file twenty times larger than the original containing no additional information whatsoever.

Opus at 24 kbps encodes speech extremely efficiently, but MP3 is a much older design and needs three to four times the bitrate to encode the same speech without adding audible warble around consonants. So the useful range is 96–128 kbps: enough that the MP3 encoder is not the bottleneck, and not so much that you are storing silence at high fidelity. Push to 192 kbps only if the recording contains music.

A one-minute WhatsApp note is around 150 KB as Opus and about 950 KB as a 128 kbps MP3. That seven-fold growth is normal and unavoidable — it is the cost of the older format, not a fault in the conversion.

Converting a whole chat export

Unzip the export, select every audio file, and the queue works through them in order. Fifty files at a time, a progress row each, and one ZIP at the end. This is precisely the workload upload converters handle worst — dozens of small files, each requiring its own round trip, all of them counting against a free-tier meter.

These recordings should not be uploaded

Voice notes are the most personal audio most people own: family conversations, medical appointments, arguments, legal discussions, evidence. Every upload-based converter is asking you to send that to an unknown server and trust a retention policy you cannot verify.

There is no server behind this page. The Opus decoder and the LAME MP3 encoder are both compiled into a WebAssembly build of FFmpeg that is served from this site and runs inside your browser tab. The site’s Content-Security-Policy lists no third-party origin, so there is nowhere for a file to be sent even by accident. Open your browser’s Network tab and convert something — after the one-time engine download there are no requests at all. Switching off Wi-Fi and converting anyway is the same proof, faster.

Related tools

Older phone recordings and call captures are usually AMR — see AMR to MP3. To merge a run of voice notes into a single file, use the audio joiner, which can trim each one on its waveform first. To make a new recording in the first place, there is the voice recorder.