What the noise reduction tool does
The cleaner runs DeepFilterNet on the uploaded recording to reduce background noise around speech. It is designed for spoken audio, such as interviews, voice notes and recorded conversations.
Noise reduction cannot restore clipped, missing or heavily distorted speech. Results depend on the recording and noise; keep the original and listen to the output before using it.
How to clean an audio recording
- Choose a supported audio or video file.
- Select Clean audio.
- Wait while the server processes the recording, then download the cleaned WAV.
The tool extracts audio from video uploads. You do not need to install software or create an account.
Supported files and free-use limits
Accepted formats: AAC, FLAC, M4A, MKV, MOV, MP3, MP4, OGA, OGG, OPUS, WAV, WebM and WMA. Each file can be up to 32 MB and 5 minutes long. One conversion runs at a time to share the CPU server fairly.
Privacy during processing
Uploads are processed in a temporary server folder. The input and output files are deleted after the download response finishes. Recordings are not used to train a model. Read the privacy details.
Choose a recording and start processing directly from your browser.
Open the free toolFrequently asked questions
Is this audio noise reduction free?
Yes. There is no account, subscription, advertising or paid audio API. The service has a shared one-job-at-a-time limit.
Does it work on music?
The tool is intended to reduce background noise around speech, not to master music or separate every instrument.
Which output do I get?
Noise reduction returns a WAV audio file.
Are results guaranteed to sound perfect?
No. Strong noise, clipping, reverberation and recording quality can affect the result. Always listen to the processed audio.