Standard Retrieval-Augmented Generation (RAG) pipelines are built exclusively for textual data. They parse markdown files, split paragraphs into token chunks, calculate dense text embeddings, and index them into stores like PostgreSQL with pgvector.…