Add support for images using GPT4-vision

[jacobSingh]

I was thinking of a cool workflow: Take a screenshot, copy to clipboard, then ask a question about that screenshot.
This would required the GPT4-vision model.

Maybe the script needs to be modified to allow the model to be passed in. Architecturally speaking, I would separate out the "request building" part from the "setting up the request part"...