📖 Build & query RAG system with Google Drive, OpenAI GPT-4o-mini, and Pinecone
⚡ 5,058 views · 📖 Internal Wiki & Knowledge Base
Description
🔍 What This Workflow Does
This RAG Pipeline in n8n automates document ingestion from Google Drive, vectorizes it using OpenAI embeddings, stores it in Pinecone, and enables chat-based retrieval using LangChain agents.
Main Functions:
📂 Auto-detects new files uploaded to a specific Google Drive folder. 🧠 Converts the file into embeddings using OpenAI. 📦 Stores them in a Pinecone vector database. 💬 Allows a user to query the knowledge base through a chat interface. 🤖 Uses a GPT-4o-mini model with LangChain to generate intelligent responses using retrieved context. ⚙️ Setup Instructions
- Connect Accounts Ensure these services are connected in n8n:
✅ Google Drive (OAuth2) ✅ OpenAI ✅ Pinecone You can do this in n8n > Credentials > New and use the matching names from the file:
Google Drive: “Google Drive account 2” OpenAI: “OpenAi success” Pinecone: “PineconeApi account 2” 2. Folder Setup Upload your documents to this folder in Google Drive:
📁 Power Folder
The workflow is triggered every minute when a new file is uploaded.
- Workflow Overview A. File Ingestion Path
Google Drive Trigger — detects new file. Google Drive (Download) — downloads the new file. Recursive Text Splitter — splits text into chunks. Default Data Loader — loads content as LangChain documents. OpenAI Embeddings — converts text chunks into embeddings. Pinecone Vector Store — stores them in “ragfile” index. B. Chat Retrieval Path
When chat message received — AI Agent — LangChain agent managing tools. OpenAI Chat Model (GPT-4o-mini) — generates replies. Pinecone Vector Store (retrieval) — retrieves matching content. Embeddings OpenAI1 — helps match queries to document chunks.
🔗 Nodes Used
Google Drive, Google Drive Trigger, AI Agent, Embeddings OpenAI, OpenAI Chat Model, Recursive Character Text Splitter
📥 Import
Download workflow.json and import into n8n:
Workflow menu → Import from File