n8n Workflow: Build RAG with Google Drive, Qdrant & Gemini

Build a RAG system that stays up to date—using Google Drive, Qdrant, and Google Gemini

This n8n workflow automates Retrieval-Augmented Generation (RAG) by indexing documents from Google Drive into Qdrant and powering question answering with Google Gemini. It also supports full or incremental updates, including replacing the vector data for a single file without reprocessing your entire collection.

What this workflow does

• Sets up (or resets) your Qdrant collection to store document embeddings, including configuration such as cosine distance metrics.
• Pulls files from a specified Google Drive folder, then downloads and processes each document.
• Extracts, chunks, and embeds text so the content can be searched effectively in Qdrant.
• Creates or updates vectors in Qdrant for new or changed content—supporting full indexing as well as incremental updates.
• Enables single file updates so you can replace a document’s representation in Qdrant while avoiding unnecessary reprocessing.
• Connects to a chatbot Q&A flow using Google Gemini to answer questions with retrieval from your vector database.

Use cases

• Keep a company knowledge base current when Google Drive documents are revised.
• Update only one policy, spec, or manual page in Qdrant after edits—without rebuilding everything.
• Support internal customer support bots that answer using the latest Google Drive content.
• Maintain a scalable RAG pipeline for SaaS operations where documents change frequently.

Technical details

• Document source: Google Drive
• Vector store: Qdrant
• Chatbot / QA integration: Google Gemini
• n8n nodes used: set, wait, sticky note, google drive, http request, manual trigger

If you want an n8n automation workflow for RAG with Google Drive → Qdrant → Gemini and reliable single-file vector updates, this is built for you.