I'm working in a large development organization and we have an internal proxy URL to access OpenAI models. I'd like to try using Kit MCP with AI assistants, but for that to work in our org I'd need to configure the OpenAI base_url to point to our proxy. Would you consider exposing the base_url parameter to configuration for OpenAI? I believe that this would be all that would be needed to allow us to try using Kit's MCP server internally.