Following up to the latest discussions on Hackernews together with this awesome community we pushed several updates to Github. Botium Speech Processing is now using way better pre-trained modesl from the Zamia Speech project and has an option to watch a file system folder for running text-to-speech and speech-to-text recipies.
https://github.com/codeforequity-at/botium-speech-processing