Exposing OpenAI API of a Google API endpoint
--port [PORT]
Support endpoint: IP:18788(--port)/v1/chat/completions
Support thinking model: gemini-2.0-flash-thinking-*
Special mode: Append "-no-thought-process" to model name to not relay its thought process, only the result