Vector Store (Vector Database)

A database optimized for storing and searching embeddings. When documents are processed for a RAG pipeline, they get converted into embeddings and stored in a vector database. At query time, the user's question also gets converted to an embedding, and the database finds the stored embeddings that are most similar, like a librarian who organizes books by vibes rather than Dewey Decimal. Examples include Pinecone, Weaviate, Qdrant, Milvus, and ChromaDB.

Why it matters for writers: The choice of vector store affects what metadata you can attach to documents, how fast retrieval is, and what filtering capabilities are available. If you're building or configuring a RAG system, understanding vector store capabilities helps you design better content ingestion pipelines. Also, expect to explain this concept roughly once a week to someone who asks "can't we just use a regular database?"

Related terms: Embedding · Similarity Search · Retrieval-Augmented Generation