Multimodal inputs #23
Description
Buttons, radio button, checkbox -- could they just send an utterance? or send a "button press"?
Camera input, for example biometrics. How does that work?
We could use the mime type to tell the agent that it's getting an image
encode image or audio as base 64 ASCII to send to server
Could WebRTC be helpful?
Does audio also need to specify an encoding along with the mime type?