Support for Ollama servers? #64
Description
Hi all,
I continue to be astonished at the speed and quality of development here, thanks to everyone involved for something which has made life better and easier. I thought I would ask, if I may, for a feature which might be more useful as time passes. Ollama
https://ollama.com/
is a method through which LLMs can be run on the local intranet. There is an accessible option for these, though at an early stage, at
https://github.com/chigkim/VOLlama/
Would it be possible for the add-on to support sending data to an Ollama server/model? It would be particularly nice to be able to send NVDA's objects to the model as well, of course, as allowing work with the local models through the central dialogue. Thanks for looking into whether this would be possible.