If checked, the application will connect to a running LM Studio instance for all model interactions. If unchecked, it will use local GGUF models specified in your .env file.