AI Document Q&A Agent: Chat with Your Files using Pinecone & OpenAI

Unlock Instant Document Insights with this AI Agent

This n8n workflow acts as an AI Agent, transforming your documents into an interactive knowledge base. It fetches a specified file from Google Drive, intelligently processes its content using text splitting techniques, generates vector embeddings via OpenAI's models, and stores them in a Pinecone vector database. Once your document is indexed, this AI Agent allows you to 'chat' with it: ask natural language questions and receive contextually relevant answers, synthesized by an OpenAI language model using Retrieval Augmented Generation (RAG). This effectively turns static documents into dynamic, queryable knowledge sources.

Key Features & Benefits

• Automated Document Indexing: Fetches files from Google Drive and automatically processes them for Q&A.
• AI-Powered Q&A: Leverages OpenAI's embedding and chat models for intelligent information retrieval and answer generation.
• Vector Search with Pinecone: Utilizes Pinecone for efficient similarity search over your document content.
• Interactive Chat Interface: Allows users to ask natural language questions and receive answers directly within n8n.
• Customizable Knowledge Base: Easily adapt to use your own documents from Google Drive.
• Streamlined Information Access: Quickly find information within large documents without manual searching.

• An n8n instance (Cloud or self-hosted).
• OpenAI API Key with access to embedding models (e.g., text-embedding-ada-002) and chat models (e.g., gpt-3.5-turbo or gpt-4).
• Pinecone API Key and Environment details.
• A Pinecone index pre-created with 1536 dimensions (if using text-embedding-ada-002).
• Google Drive OAuth2 credentials configured in your n8n instance.
• The URL of the Google Drive file you wish to process.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Configure Google Drive Access: In the 'Google Drive' node, ensure your Google Drive OAuth2 credentials are selected. The 'Set Google Drive file URL' node pre-fills a sample file URL; update this node or the 'Google Drive' node's 'File ID' parameter (set to URL mode) with your target file's URL.
4. Configure OpenAI Credentials:
   • In the 'Embeddings OpenAI' node (for indexing), select or add your OpenAI API credentials.
   • In the 'Embeddings OpenAI2' node (for querying), select or add the same OpenAI API credentials.
   • In the 'OpenAI Chat Model' node, select or add your OpenAI API credentials.
5. Configure Pinecone Integration:
   • In the 'Insert into Pinecone vector store' node: Select or add your Pinecone API credentials, specify your Pinecone Environment, and choose your Pinecone Index name (e.g., 'test-index'). Ensure 'Clear Namespace' is checked if you want to re-index into a clean namespace.
   • In the 'Read Pinecone Vector Store' node: Select or add your Pinecone API credentials, specify your Pinecone Environment, and choose the same Pinecone Index name.
6. Verify Pinecone Index: Ensure the Pinecone index you're using has 1536 dimensions, compatible with OpenAI's text-embedding-ada-002 model used by default.
7. (Optional) Adjust Text Splitting: Modify chunkSize and chunkOverlap in the 'Recursive Character Text Splitter' node if needed for your document's structure.
8. Load Your Document: Click the 'Test Workflow' button (or the play icon on the 'When clicking 'Test Workflow' button' node). This triggers the workflow to download, process, embed, and store your document in Pinecone.
9. Chat with Your Document: Once the loading process is complete (check the execution log), click the 'Chat' button at the bottom of the canvas (associated with the 'When clicking 'Chat' button below' trigger). Enter your question in the prompt.
10. Activate the workflow for persistent use.