Effortless Vector Database Loading from Google Drive with n8n

This n8n workflow is designed to bridge the gap between your cloud storage and your vectorized data infrastructure. It begins by automatically scanning a designated Google Drive folder for new or updated files. Once identified, each file is downloaded, its content is extracted based on its MIME type (supporting PDFs and plain text), and then intelligently segmented into smaller, digestible chunks using the Recursive Character Text Splitter. For each chunk, OpenAI embeddings are generated, leveraging the text-embedding-3-small model. Finally, these vectorized representations are efficiently inserted into a PostgreSQL database with the PGVector extension, specifically into the n8n_vectors_wfs table within the n8n_wfs collection. Upon successful vectorization, the original file is moved to a 'vectorized' folder in Google Drive, keeping your source organized.