Noise Removal (for Speech)
Reduce background noise in a recording and save a clean WAV.
Photos, audio and text are processed only in this browser and are never sent to a server or stored, including photos that contain personal information. Only the AI model files are downloaded the first time you run a tool.
AI results can be wrong. Do not use them for important decisions.
Made for speech. Listen to the result before using it.
How to use
- Upload an audio or video file.
- Set the strength. 100 is fully processed; lower values mix in the original sound.
- Press Run and wait.
- Listen to it and save the WAV.
FAQ
Does it work for music?
It is made for speech. On music it can sound muffled or cut instruments. We do not recommend it for music.
What noise does it handle well?
Steady noise such as fans, wind and air conditioners. Sudden sounds and other people talking are hard to remove.
How do I choose the strength?
If the result sounds unnatural, lower the strength to about 70-80 to mix in the original.
What is the output quality?
A 48 kHz mono WAV. Stereo is mixed down to mono, and for video only the sound is saved.
Models and licenses used by this tool
- RNNoise (@jitsi/rnnoise-wasm)BSD-3-Clause / Apache-2.0ModelRNNoise © Jean-Marc Valin, Mozilla, Xiph.Org Foundation, Amazon (BSD-3-Clause). WebAssembly build: Jitsi (Apache-2.0).2026-11-11
Licenses were checked against the original model cards and repositories (checked 2026-11-11). They are listed here to give the required credit. Sources & licenses