I’m currently working on a voice operated personal assistant based on python.  So far it uses Weather Underground and Wolfram Aplha APIs.  Using speech_recognition.py to recognize voice via Google’s speech recognition.  Although I’d like to update this to use cmusphinx for offline recognition of the prompt phrase.  I shall be updating this in the future with additional information concerning it’s dependencies and possibly some code snippets.

Update: I bought two Google Homes and put this project down.