This is impressive, well done. I didn't find it on the roadmap, but do you plan to add support for local LLM models like llama3? I think it should be possible t

This is impressive, well done. I didn't find it on the roadmap, but do you plan to add support for local LLM models like llama3? I think it should be possible to launch a local server from LM studio or similar and try to run it even now, but native support would be much better. Or maybe you tried it, but it's not good enough?