The MCP RAG Server is a protocol-based server designed to empower Large Language Models (LLMs) with retrieval capabilities. It indexes documents in multiple formats such as text, markdown, JSON, and CSV, and splits them into chunks based on configurable size. The server processes these chunks through embedding APIs, storing vectors locally in an efficient SQLite-based vector store. During inference, it embeds user queries and retrieves the most relevant text chunks, providing context-aware responses. This setup enhances the accuracy and relevance of LLM outputs when working with document collections, making it ideal for knowledge bases, document search, and enterprise data questions. The server exposes various tools and APIs for document management and querying, supporting seamless integration with custom clients and workflows.