Google Drive to Pinecone Vector Store with Context-Aware Chunking using Gemini

Overview

This n8n workflow automates the process of ingesting documents from Google Drive into a Pinecone vector database for Retrieval Augmented Generation (RAG) applications. It begins by downloading a specified Google Drive document. The document's text is then split into manageable sections. Crucially, an AI agent (using OpenRouter and Gemini) is employed to generate context-aware summaries for each section. This enhanced text, combining the original chunk with its context, is then embedded using Google Gemini and stored in Pinecone. This approach improves the accuracy and relevance of retrieved information in RAG systems by providing richer context for each data chunk.