This project is an LLM invocation platform designed for efficient content generation in specific environments, considering resource constraints and data privacy needs. It integrates multiple locally deployable LLMs (chatGLM-6B, chatRWKV, llama, InternLM, etc.) and supports OpenAI API access. The platform also includes a knowledge base supporting local offline vector databases, local search engines, and online search. A JavaScript-based Auto scripting system allows custom plugin development for custom conversation flows and dynamic model switching.