Instructions for local LLMs unclear for newbies

[kolenyo2099]

Hi all,

I was trying to make use of the local LLM connection feature, but it took me a while to remember that, when connecting to local host from docker, one should switch http://127.0.0.1 for http://host.docker.internal, and thus, when using ollama or lmstudio, with something like http://host.docker.internal:1234/v1.

I thought perhaps there's a fast way of communicating this in the UI.

Cheers,

g.

[sal-uva]

Well spotted! I think we can detect whether 4CAT is running in Docker, and automatically update the default URL to host.docker.internal if nothing is inserted.