Audio tools that run in your browser

This is the audio half of the catalogue on one page: the editing operations, the format converters and the analysers, without the video tools they normally sit beside. Every one of them decodes your file with the browser's own Web Audio engine, works on the samples in memory, and encodes the result back out in the tab.

Fix a recording. The normalizer raises a quiet file to a target level without clipping the peaks — the standard answer to one voice being half the volume of the next. Noise reduction removes the steady hiss, hum or fan a microphone picked up. The silence remover and the pause remover cut dead air out of speech: the first by threshold across the whole file, the second tuned for the long gaps in a spoken take. The equalizer gives ten bands with presets, fades handle the top and tail, and the mixer lays several tracks over each other.

Change the shape of a file. The trimmer cuts to a selection against a waveform, and the joiner concatenates several files into one track with an optional crossfade at each seam. Speed changes the tempo; the pitch changer moves the key without touching the tempo, which is the one people search for and the one most simple tools get wrong. The waveform generator exports a picture of the audio for a thumbnail or a podcast page. The tag editor writes ID3 title, artist, album and cover art back into an MP3.

Change the format. The general converter moves between MP3, WAV, OGG, FLAC and AAC, and the dedicated pairs cover the awkward inputs: M4A from a phone, OPUS from Discord or WhatsApp, AC3 from a film rip, APE from a lossless archive, WMA from an old Windows library, AIFF from a Mac session. Audio can be pulled out of a video, or combined with a still image to make a video a platform will accept.

Measure it. The BPM and key detector reads tempo and musical key out of a track, and speech-to-text dictates live through the browser's own recognition engine.

How to choose an order of operations: trim first, then repair, then level, then convert. Cutting before you denoise means less material to process; normalising before you convert means the encoder is working on the final dynamics; converting last means you only encode once. Going lossy twice — MP3 in, MP3 out — costs quality every time, so keep a WAV or FLAC master if there will be more edits. And for a codec choice: WAV and FLAC for anything you will edit again, MP3 or AAC for distribution, OPUS for voice at small sizes.

Nothing is uploaded. `decodeAudioData` turns your file into raw samples inside the page, the edit runs on those samples, and the encoder writes a new blob your browser hands straight to the download — no request carries the audio, so there is no queue, no processing limit and no file sitting on someone else's disk afterwards. The practical ceiling is memory: an hour-long WAV decoded to floating-point samples is large, so very long files are better cut into parts first. Video editing, subtitles and burned-in captions live on the Audio & Video hub, which is the same catalogue with the video tools included.