Browse projects/Audio and music/Remove background noise from audio/DeepFilterNet
rikorose/deepfilternet A tool that works on your files
Remove background noise from a voice recording
- Open source
- Runs on your own computer
- Windows, Mac or Linux
- Tested 4 Nov 2026
You bring
- a speech recording in a format FFmpeg reads
You get
- a denoised mono WAV file at 48 kHz
What can it reach?
- The internet. Downloads the official CPU executable with its embedded model from the upstream release on first use. The download container cannot read your recording.
How was it tested?
Run on clean Intel and AMD and Apple silicon and ARM machines on 4 Nov 2026. Its check passed on a known sample and rejected deliberately broken results.
This checks the job this page describes, not every feature of the original project.
Project, licence and recipe files
rikorose/deepfilternet on GitHub
Licence. MIT OR Apache-2.0 (source); the model is embedded in the upstream executable. Neither the executable nor its weights are redistributed in this image.
Recipe revision 8bcd16ba9971, from 978576aa8400.
Audio and music
Transcribe, separate, clean up and voice your audio.
- demucs adefossez
Split a song into vocals, drums, bass and the remaining instruments
You get vocals
- spleeter deezer
Split a song into a vocal track and an instrumental backing track
You get the isolated singing
- kokoro hexgrad
Read English text aloud in a chosen voice and at a chosen speed
You get a spoken recording of your text



