RAG Chatbot: Google Drive Q&A via Telegram & GPT-4-mini

This workflow implements a Retrieval-Augmented Generation (RAG) system that enables users to upload documents to a monitored Google Drive folder. Upon upload, it automatically downloads the files, processes them into vector embeddings using OpenAI, and stores them in an in-memory vector store for efficient retrieval. This setup transforms static documents into a dynamic knowledge base accessible via natural language queries.

When users send questions through the integrated Telegram bot, an AI agent powered by GPT-4-mini and LangChain performs similarity searches on the vector store, retrieves relevant document chunks, and generates accurate, context-aware responses with citations. Short-term conversation memory ensures coherent multi-turn interactions, enhancing user experience.

Key benefits include seamless automation of document ingestion and querying, reducing manual research time, and providing instant insights from proprietary data. Ideal use cases: internal knowledge bases for teams, student study assistants with uploaded materials, business report summarization, and customer support for FAQ handling via chat.