Don’t remember exactly which guide I followed, but I used the build from this repo, and the instructions looks plausible:

https://github.com/werman/noise-suppression-for-voice#pipewi...