This Node.JS server controls Misty via her REST API. It communicates with external services to do speech recognition, response generation and text-to-speech. It also uses Misty’s built-in face recognition to try and look people in the eye while talking. More info in the README.
Also, checkout this nifty little library on top of which
misty-interact was built. It makes sending requests to Misty and binding to websocket events pretty easy. It also wraps websocket event streams in RxJS observables so you can easily map, group, filter, combine and transform all kinds of event streams coming from Misty’s sensors.