Ask HN: Any libraries to animate AI avatars based on sound?

I would like for people to submit some photos from which we can make a 3D model, or even better to have their face mapped as they are saying certain things. Then we would generate their avatar for when they are talking on WebRTC, so they don’t have to be present on video. I realize the potential for abuse if the models were leaked or someone in our company was bribed, but does the technology exist today? 1. From photos on the web 2. From videos on the web 3. From interactively registering via webcam and face mapping