How difficult would it be to add in a feature where you could feed the program say video.mp4, and have it censor it frame by frame, then pump out a video_censored.mp4 version? You already have ffmpeg as part of the program so I assume splitting audio, censoring and recombining could all be handled with that. At the moment I do something similar with ffmpeg but it applies one kind of blurring, to the whole video rather than specific filters to detected parts.