Chat with Google Drive Documents using GPT, Pinecone, and RAG

Description

📌 Short Overview

Automatically sync files from Google Drive into a searchable AI knowledge base with Pinecone, and answer user queries using GPT-4o with conversational memory.

⸻

🛠️ Workflow Usage Steps

1. Watch Google Drive for file changes
Trigger the workflow when a new file is uploaded or an existing file is updated in a specific Google Drive folder.
2. Download and process the file
Retrieve the file, split it into smaller text chunks with a Recursive Character Text Splitter, and generate vector embeddings using OpenAI.
3. Store embeddings in Pinecone
Save the embeddings in a Pinecone vector database to keep your knowledge base continuously updated and searchable.
4. Search context for chat queries
When a user asks a question, query Pinecone for relevant context, combine results with conversational memory, and process them with GPT-4o.
5. Respond with AI-powered answers
Provide a concise response (100–200 words) that blends knowledge from your documents with the conversation history.

⸻

✅ Use Cases
• Keep a live, AI-ready knowledge base from your Google Drive files.
• Enable team members to query company documents instantly.
• Build a personal assistant that stays up to date with your latest uploads.

⚙️ Setup Steps
Google Drive
• Create a Google Cloud project.
• Enable the Google Drive API.
• Generate OAuth credentials and connect them in n8n.
OpenAI
• Sign up at OpenAI.
• Copy your API key from the dashboard.
• Add it to n8n under Credentials → OpenAI API.
Pinecone
• Create an account at Pinecone.
• Create a new index (e.g., docs-embeddings).
• Copy your API key and environment, then add them to n8n under Credentials → Pinecone API.
Workflow Configuration
• Import this workflow into your n8n instance.
• Select the Google Drive folder you want to monitor.
• Set the Pinecone index name in the workflow.
• Adjust chunk size / overlap in the text splitter if needed.
Test the Workflow
• Upload a new document to your Google Drive folder.
• Run the workflow to confirm embeddings are created and stored in Pinecone.
• Ask a sample query and verify the AI returns a context-aware answer.