About TidyAudio

TidyAudio started as a private tool. I record a lot of spoken audio, and the same four chores came up every single time: cut out the fumbled sentences, get rid of the hiss the room adds, even out the volume because I drift away from the microphone, and convert the result into something that will actually play on the other end. None of those are hard problems. They just were not available together, for free, without handing my recordings to a stranger's server.

So I built the thing I wanted, used it for months, and then it seemed daft not to put it where other people could use it too. That is the entire origin story. There is no company here and no funding round: one person, one page of code, and a domain name.

The one decision everything else follows from

TidyAudio does its work in your browser, on your own machine. That single architectural choice is what makes the rest of the site's promises possible, and it is worth spelling out why, because "we respect your privacy" is a sentence anyone can type.

Most online audio tools are a thin web page in front of a server. You upload your file, their computer does the work, you download the result. That design means your recording sits, however briefly, on hardware you do not control, subject to a retention policy you have to take on faith. It also means the operator pays for every conversion, which is why those sites eventually meter you, watermark the output, or ask you to sign up.

TidyAudio has no server to upload to. The editor is delivered to your browser as a page, and from that moment your own device does the decoding, the editing, the noise filtering and the encoding. Your audio cannot leak from a server that never receives it. This is not a policy that could quietly change in a future update to the terms; it is a property of how the software is put together, and you can confirm it yourself by opening your browser's network tab and watching what does not get sent.

The happy side effect is that running this site costs me almost nothing, which is why every feature is free with no account, no watermark and no file size tier. There is no expensive per-user cost that I would eventually need to claw back from you.

What is actually inside

The editor is one self-contained HTML page. It uses the Web Audio API for decoding and playback, the canvas element for the waveform, and Web Workers so that slow jobs like noise removal run on a background thread instead of freezing the page while you wait.

Two pieces of open-source code do work browsers cannot do natively. The LAME encoder writes MP3 files, and the opencore-amr decoder reads the .amr voice memos that phones produce, a format no mainstream browser will open. Each is fetched once from a public CDN the first time you use that feature, then cached. They are program code, not your audio.

The noise reduction deserves an honest description, because this is an area where marketing usually outruns reality. It is a Wiener filter with decision-directed signal estimation: it learns what your room sounds like when nobody is speaking, then decides frequency by frequency and moment by moment how much of what it hears is noise. That is a genuinely good classical technique, and on a normal recording it removes background while leaving the voice within about a decibel of untouched. It is not an AI speech model. Tools that resynthesise a voice with a neural network can invent detail that was never recorded and can remove room reverb; this cannot, and I would rather say so plainly than let you discover it after uploading your one irreplaceable interview.

Who it is for

People with a recording and a problem, rather than audio engineers. If you already own Reaper or Audition, keep using them; they will beat this at almost everything. TidyAudio is aimed at the much larger group who recorded a lecture, an interview, a phone call, a practice session or a voice memo, need it cleaned up and converted once, and do not want to install a digital audio workstation to do it.

It is especially useful when the recording is sensitive. Medical dictation, legal calls, therapy notes, journalism, HR conversations: all the material where "just upload it to a free website" is a genuinely bad idea. That is the case the no-upload architecture was built for.

What it will never do

• Ask you to sign up. There is no account system, because there is nothing to store.
• Watermark or truncate your output. There is no paid tier being protected.
• See your audio. Not the files, not their names, not their contents. The analytics on this site count page views; they have no access to anything you open in the editor.
• Pretend to be magic. Where a limit exists, the guide states it: AMR is telephone quality and conversion cannot improve on that; the noise filter cannot remove a barking dog; a browser tab has finite memory, so hour-long files are happier on a desktop.

How it pays for itself

Modest advertising, and nothing else. There is no premium plan, no data sale, and nothing about your audio is available to sell even if I wanted to. If the ads ever become obnoxious enough to spoil the tool, that is a bug in my judgment and I would like to hear about it.

Corrections and requests

If something here is wrong, unclear, or does not work on your machine, I would rather know. The guide documents every feature in detail, and contact tells you how to reach me. Feature requests genuinely do get built: several of the tools on this site exist because they were the next thing I needed.