AI-Powered Document Search with Vector Embeddings

This n8n workflow enables intelligent document retrieval and question-answering by integrating OpenAI embeddings with a Supabase database. When a chat message is received, the workflow generates a semantic embedding of the query and searches the database for the most relevant files based on vector similarity. It filters results to ensure high relevance, aggregates matching file IDs, and fetches content along with URLs. Designed for use in knowledge management, customer support, or internal documentation, this automation helps users quickly find precise information from large collections of text or PDF files. The workflow also allows easy setup, with options to configure similarity thresholds, match counts, and credentials for seamless integration.