A lightweight proxy server to call over 100 LLM APIs using the standardized OpenAI input and output format.