Open an audio file, drag the handles on the waveform to select the part you want, preview the selection, add a fade in or out of whatever length you choose, mix to mono if you need to, and export the result as a 16-bit WAV. The browser decodes audio at the computer’s own sample rate - usually 48 kHz - so the WAV comes out at that rate, which may not be the original file’s.
It can also reduce background noise under speech - hiss, hum, fans, room tone - with RNNoise, a small open-source neural network that runs on your computer, with no limit on how much you clean. Play the selection with and without it before you download, and turn the strength down if full cleaning sounds too processed. It is made for voices: it is not meant for music, overlapping speakers or echo. For heavier noise, or the sound of a video, Remove Background Noise uses DeepFilterNet, a larger model that gives a cleaner result.
Handles MP3, WAV, OGG, FLAC and M4A. Common uses are cutting a ringtone, trimming the silence off a recording, or cleaning up a voice memo or interview recorded near a noisy fan. It is free to use, with no account and no upload limit, and it runs entirely in your browser: nothing is uploaded, and while the page stays open, a feature you have already used keeps working offline.
- Why does it export WAV rather than MP3?
- WAV is what a browser can write reliably without an extra encoder, and it is uncompressed, so the trim adds no compression loss of its own. Convert it afterwards if you need a smaller file.
- Is my recording uploaded?
- No. It is decoded and trimmed with the Web Audio API in the page. An interview, a voice memo or an unreleased track never leaves your machine.
- Will trimming reduce the quality?
- Barely. The clip is cut from a full-quality decode and is not compressed again, so there is no second round of MP3-style loss. Two things do change: the browser decodes at the computer’s own sample rate, so a 44.1 kHz file usually comes out resampled to 48 kHz, and the WAV is written as 16-bit - CD quality, but less than a 24-bit studio recording holds.
- What does mixing to mono do?
- It folds the two stereo channels into one, roughly halving the file size. Good for speech, less so for music where the stereo image matters.
- Which formats can it open?
- Whatever your browser can decode, which in practice means MP3, WAV, M4A (and iPhone M4R ringtones), OGG and FLAC on a current browser, and MKA in Chrome and Edge. The whole file is read into memory, so a file of several gigabytes, such as a long video, may be too large to open: save its sound as audio first.
- What kind of noise can the audio trimmer remove?
- Steady background noise under speech: tape or microphone hiss, mains hum, a fan or air conditioner, the general sound of a room. It uses RNNoise, a neural network trained to tell speech from noise, which runs in the page. It reduces that noise rather than always removing it completely - the louder the noise, the more is left - and in the gaps between words it goes quietest.
- Will the noise reduction work on music, or on echo?
- No. It is trained on speech, so it treats most music as noise and can strip or warble it. It cannot pull one voice out from others talking over it, and it does not remove echo or room reverberation. Where it cannot help, set the strength lower or leave it off.
- Is there a limit on how much audio I can clean?
- No. Online voice enhancers often cap how much you can process for free, because the work runs on their servers. Here it runs on your own computer, so there is nothing to meter. A long selection takes a while - a minute or two for an hour of stereo on a typical laptop - and shows its progress, and it can be cancelled.