关于
This project delivers a production-style Retrieval-Augmented Generation (RAG) system, allowing users to upload multi-page PDF documents and engage with their content via natural language questions. It integrates Google Gemini for advanced embeddings and answer generation, leveraging BigQuery Vector Search as a scalable vector database for efficient semantic search. The system employs LangGraph for deterministic orchestration of the RAG workflow and utilizes the Model Context Protocol (MCP) to expose its capabilities as a versatile tool. An interactive chat UI, built with Streamlit, provides a seamless user experience, supporting sophisticated document analysis and structured agent execution.
