Persona is a collection of different characters from the future that can be talked to and asked about their life. Each persona is given a high quality animated avatar and is connected to an LLM. Users can text or speak with the persona and see and hear the animated response.
Personas' backstories were developed using closed IP predictions about the future, work styles, lifestyles, technology that has become everyday, etc. Speech to text (STT) and Text to Speech (TTS) pipeline for handling conversation-like interaction. Text goes through a RAG look up of relevant future predictions pertinant to the persona. Facial and eye animations are programatically generated live using viseme data as well as metadata added to the generated LLM response. Body animations are a series of pre-canned skeletal animations.