This demonstrates some different technologies. It uses Unreal Engine Metahumans for rendering. I use DeepGram to turn my voice into text. I use OpenAI's ChatGPT API to turn my voice as text into dialogue for Polly. I then use Amazon Polly to turn the response back into Polly's voice. There are a lot of things to tweak still, including animation and body language (some good work by Ubisoft to incorporate and I want to try NVidia's voice to facial animation).