Audio

About these tools

Convert between common audio formats and pull the audio track out of a video, using a WebAssembly build of FFmpeg that runs in your browser. The engine is a few megabytes and downloads the first time you use one of these tools, then converts everything locally.

This is the slowest category on the site, and honestly so: audio encoding is real work, and doing it on your device rather than on a server means your machine does that work. What you get in exchange is that recordings, interviews, and voice notes are never transmitted.

The choice in audio is compression, and it is a one-way door. WAV stores every sample as recorded and costs about thirty megabytes for three minutes. FLAC compresses that by roughly half and gives every sample back exactly. MP3, AAC, Opus and Vorbis throw away sound the encoder judges inaudible and get to a few megabytes, and what they discard cannot be recovered by converting back — a WAV made from an MP3 is a large file containing MP3-quality audio.

Everything here runs through FFmpeg compiled to WebAssembly, which is the same engine most desktop converters wrap, running in the page instead of on a server. That is why the first conversion pauses to fetch the engine and later ones do not, and why an interview, a voice memo or an unreleased track is decoded and encoded on your own machine rather than uploaded to somebody's queue.

Frequently asked questions

Can I really convert audio in a browser?

Yes. FFmpeg is compiled to WebAssembly and runs inside the page. It is the same decoder and encoder set that desktop audio tools use, so the output is ordinary MP3, WAV, OGG, Opus, or FLAC that any player will open.

How do I extract audio from a video?

Use the video to audio tool, choose the video, and pick an output format. The video is decoded on your device and only the audio track is written out. Nothing about the video is uploaded, which matters for personal recordings.

Why does the first conversion take longer?

The audio engine has to download before the first conversion, because bundling several megabytes of WebAssembly into every page would slow down the whole site for people who never convert audio. After that it is cached and subsequent conversions start immediately.

Does converting MP3 to WAV improve the quality?

No, and no tool can. MP3 discarded detail when it was created, and decoding to WAV simply stores what remains in a lossless container: a bigger file with identical sound. Converting to WAV is useful for editing, not for quality.

Which format should I choose?

MP3 for maximum compatibility, Opus for the best quality per megabyte on modern devices, FLAC when you want lossless archiving, and WAV when a program insists on uncompressed audio. Each tool lists the bitrate presets it offers.

Can I edit, trim, or normalise audio here?

No. These are conversion tools, not an editor. Trimming, mixing, normalising, and noise removal are a different product, and adding a poor version of one would not serve anybody.

How large a file can I convert?

Up to 250 MB. The practical limit is memory and patience rather than the file size itself: a long recording takes real time to encode, and the page tells you what it is doing throughout.

Does converting change the volume or the length?

Neither. The audio is decoded and re-encoded sample for sample, so the track comes out the same length and at the same level it went in. Nothing is normalised, trimmed or faded, because a conversion that quietly changed what you gave it would be the wrong kind of helpful.

What happens to artist and album tags?

They are not carried across. The audio is decoded to samples and encoded again, and the tag block does not survive that round trip, so the result arrives with the title, artist and cover art blank. If your library depends on them, tag the converted file afterwards, or keep the original alongside it.

Is anything about my recording sent anywhere?

No. FFmpeg is compiled to WebAssembly and runs inside the page, so the recording is decoded and encoded on your own machine. The engine itself is a download the first time, and that is the only request involved: it fetches the converter, never your audio. An interview or a voice memo stays where it is.