Annual Membership 50% OFF - Limited Time!

AI Noise Reducer - Remove Background Noise from Audio

Upload a voice memo, podcast, interview or lecture and remove steady background noise such as fans, AC hum, computer noise, traffic and room hiss. The RNNoise neural network runs in your browser, so your audio never leaves your device. Compare before and after, then download MP3, WAV or M4A.

How to Remove Background Noise from Audio

From a noisy recording to clear speech in four steps

01
Upload your recording
Drop an MP3, WAV, M4A, AAC, OGG or FLAC file. Any file your browser can decode works, up to 10 minutes of stereo or 20 minutes of mono audio.
For a long recording, cut out the part you need first with the audio trimmer.
02
Set the strength
At 100% the tool removes as much noise as it can. Lower the strength to keep some room tone if the voice starts to sound thin or watery.
Start at 100%, then try 60–80% if the result sounds over-processed.
03
Choose the output and remove noise
Pick MP3, WAV or M4A (128, 192 or 320 kbps; WAV is 16-bit) and click Remove noise. A typical voice memo is done in a few seconds.
The output is 48 kHz, the native sample rate of RNNoise.
04
Compare and download
Play the original and the cleaned version in the two players, then download the result.
If the voice is quiet afterwards, run the file through the audio normalizer.

Noise Reducer Features

Speech-focused noise removal that runs on your own device

01
Neural noise suppression
RNNoise, a recurrent neural network trained on speech by Jean-Marc Valin (Xiph/Mozilla), separates the voice from steady background noise frame by frame.
02
Private, in-browser processing
The model is compiled to WebAssembly and runs in a Web Worker on your device. Your audio is never uploaded and no account is needed.
03
Faster than real time
Most voice memos are cleaned in a few seconds. The first run downloads a ~0.1 MB model; MP3 and M4A output also loads a ~30 MB encoder once, which is then cached.
04
Adjustable strength
Set the strength from 0 to 100%. Lower values keep some room tone so the voice does not sound thin or over-processed.
05
Before-and-after comparison
Two players let you listen to the original and the cleaned audio before you download anything.
06
MP3, WAV and M4A export
Open MP3, WAV, M4A, AAC, OGG or FLAC files and save as MP3 or M4A at 128, 192 or 320 kbps, or as 16-bit WAV at 48 kHz.

Noise Reducer FAQ

Questions about removing background noise from audio

[email protected]

Steady background noise behind a voice: fans, air-conditioning hum, computer noise, traffic, room hiss and light keyboard noise. It works best on speech, such as voice memos, podcasts, interviews, lectures, voice-overs and video call recordings.

It is not made for music. RNNoise was trained on speech, so on a song it can remove instruments or make the sound watery. Use it on spoken recordings.

No. It keeps speech, so people talking in the background usually stay, and it does not remove background music or echo and reverb. Very loud noise and sudden bangs are reduced less, and it cannot repair clipped (distorted) audio.

No. RNNoise is compiled to WebAssembly and runs in a Web Worker in your browser. The file is processed on your device and you do not need an account.

You can open MP3, WAV, M4A, AAC, OGG and FLAC, as long as your browser can decode the file. You can save the result as MP3 or M4A at 128, 192 or 320 kbps, or as 16-bit WAV. The audio is processed at 48 kHz, the native rate of RNNoise, so the output is 48 kHz.

Up to 10 minutes of stereo or 20 minutes of mono audio per file. Processing is much faster than real time: a typical voice memo takes a few seconds. The first use downloads a ~0.1 MB model, and MP3 or M4A output downloads a ~30 MB encoder once; both are cached.

Lower the strength, for example to 60–80%. This keeps a little room tone and makes the voice sound more natural. Very noisy recordings may still keep some artifacts.

Yes. The GenMusic voice recorder has a one-click Remove background noise switch that uses the same RNNoise engine.

EXPLORE AI SONG MAKER TOOLS

Multiple Ways to Create AI Music

Creating music has never been easier. Whether you have a story to tell, lyrics waiting for a melody, or just want to experiment with sound, our AI tools help you bring your musical ideas to life in minutes.

AI Music Generator

Describe your vision, get a complete song.

Generate AI Music

Lyrics to Music

Text to Rap Music

AI Instrumental Generator

Online MIDI Editor

Key & BPM Finder

Text to Music

AI Lyrics Generator

Audio Converter

Rhythm Game

Tap Rhythm Game

Music Visualizer

BPM Tapper

Online Instrument Tuner

Online Metronome

Voice Recorder with Effects

AI Noise Reducer

Circle of Fifths

Note Frequency Calculator

Audio to MIDI

Audio to Jianpu

MIDI to MP3

Music Terminology

Chord Progression Generator

Ear Training

AI Pop Music Generator

AI Rock Music Generator

AI Jazz Music Generator

AI EDM Music Generator

AI R&B Music Generator

AI Blues Music Generator

AI Folk Music Generator

AI Metal Music Generator

AI Punk Music Generator

AI Funk Music Generator

AI Techno Music Generator

AI House Music Generator

AI Trap Music Generator

AI Ambient Music Generator

AI K-Pop Music Generator