n8n RAG Chatbot Workflow: Google Drive → Pinecone + OpenAI Webhook

Turn your Google Drive documents into a live RAG chatbot—powered by Pinecone and OpenAI Webhook

This n8n RAG chatbot workflow automatically syncs updated files from a Google Drive folder into a Pinecone vector index using OpenAI embeddings, then provides a webhook chat endpoint where an OpenAI agent answers questions by retrieving the most relevant context from Pinecone.

What this workflow does

• Monitors your Google Drive by triggering every minute when a file in the selected folder is updated.
• Extracts and parses content: reads file metadata, downloads the updated document from Google Drive, and parses it into text.
• Creates embeddings and upserts to Pinecone: splits text into chunks, generates OpenAI embeddings per chunk, and inserts them into your configured Pinecone index under the selected namespace.
• Exposes a POST webhook for chat: receives user questions and optionally a session ID to maintain conversation memory.
• Runs RAG with Pinecone retrieval: uses an OpenAI chat model with a Pinecone retrieval tool to search relevant chunks and answer based strictly on retrieved context.
• Returns responses to the webhook caller.

Use cases

• Customer support teams answering questions from knowledge stored in Google Drive policies, FAQs, or manuals.
• SaaS operators building an internal “document assistant” that stays current as files change.
• Automation engineers creating a production-ready RAG pipeline with Pinecone + OpenAI via n8n webhooks.

Technical details

• Triggers: Google Drive (watch folder; checks every minute)
• Processing: downloads file, parses to text, chunks content
• Embeddings: OpenAI embeddings for chunk vectors
• Vector storage: Pinecone index + namespace
• Chat endpoint: n8n webhook + OpenAI chat model with conversation memory and Pinecone retrieval tool
• Nodes used: set, webhook, sticky note, google drive, n8n-nodes-langchainagent, respond to webhook

Setup is straightforward: connect Google Drive credentials, add OpenAI credentials for embeddings and chat, configure Pinecone index/namespace, activate the workflow, then use the generated webhook URL to send questions.