This workflow builds a Retrieval-Augmented Generation (RAG) chatbot that integrates with Google Drive, Qdrant Vector Database, and Google Gemini AI. It allows you to upload documents from a Google Drive folder, process them into embeddings, and store them in Qdrant for fast retrieval. Users can then chat with their documents using an intelligent conversational interface powered by Gemini, with full chat history stored in Google Docs.
Ideal for teams and organizations that need a document-aware AI assistant capable of retrieving precise, context-rich answers from large document repositories.
Key Features
- 📂 Google Drive Integration
- Automatically fetches files from a specified folder.
- Supports file downloads and text extraction.
- 🧩 Document Processing & Embeddings
- Splits documents into chunks for better retrieval.
- Extracts metadata (themes, keywords, insights).
- Creates vector embeddings using OpenAI.
- 🗄️ Vector Storage with Qdrant
- Stores processed documents in a scalable vector database.
- Enables fast semantic similarity search for RAG.
- Includes human-verified delete operations to prevent accidental data loss.
- 💬 Intelligent Chat Interface
- Powered by Google Gemini AI (with optional OpenAI fallback).
- Retrieval-Augmented responses ensure context-aware answers.
- Supports continuous chat memory and contextual history.
- Saves chat logs into Google Docs for long-term reference.
- 🔔 Human-in-the-Loop Safety
- Telegram bot integration for approvals and notifications.
- Confirmation required before destructive actions (e.g., deleting vectors).
- ⚙️ Customizable Workflow
- Easily adjust folder ID, Qdrant collection name, and model parameters.
- Supports batch processing and real-time chat queries.
