(Feature Request) Allow setting a custom OpenAI API endpoint to experiment with locally hosted LLMs

[sigmondkukla]

Some open-source, locally run LLMs have the capability to emulate key features of the OpenAI API to "appear" to an application as GPT. Can you add support for a custom API endpoint to experiment with this and see if the performance of something like WizardLM or Falcon may be up to the task of generating commands?
I would then envision an input field in the Settings window with the default OpenAI API endpoint URL replaceable with your own.

[FireCubeStudios]

Would i tbe possible to test this "see if the performance of something like WizardLM or Falcon may be up to the task of generating commands" in a seperate prototype app first?

I have another WIP all in one LLM app which could support custom LLM but thats only for that app. I plan to stick to GPT 3/4 for Run

[EFLKumo]

Also need to customize the endpoint so I can call the OpenAI API from a third party instead of using the official one (they have rate limits, etc. and I haven't used it in a while)