Adding additional LLMs

[malexw]

Hey folks, I saw in the readme that you're interested in adding additional LLMs once you have confidence that they'll work well. Is there a method that you would use to determine if an LLM is working? I'm interested in adding support for Cohere's Command and I'm curious how that would align with your plans.

[malexw]

I've created a PR with a demo of Command as an LLM backend: #127