Frequently asked questions
Is my audio uploaded?
No. The audio is decoded and cleaned entirely in your browser on your device, so nothing is uploaded and it works offline.
What kind of noise does it remove?
Steady background noise such as hiss, mains hum, fans and air conditioning on voice recordings. It removes low rumble and lowers the background during quiet gaps so speech stands out more.
What it cannot do
It is not a magic filter for loud, sudden or changing noise, background music, or several voices at once. For those, results will be limited.
Which formats can I use?
Any audio your browser can open, including MP3, WAV, M4A and OGG. The cleaned file is saved as a WAV, which is lossless and plays everywhere.
How do I get the best result?
Start on Medium and compare with the original using the two players. Use Light if speech starts to sound processed, or Strong for noisier recordings. Keep Remove rumble on for hum and low background.
What are the two engines?
The Quick engine uses RNNoise, a small neural network built for voice, and runs instantly on your device with no download. The ffmpeg engine uses spectral denoising and downloads a built-in engine (about 31 MB) the first time. Quick is the best choice for most voice recordings.