So basically allow you to point the LLM to a hosted model on your local network if I understand correctly?