A vector database stores numeric embeddings of text, images, or events so similar items can be retrieved even when wording differs. It is the retrieval layer inside most retrieval-augmented generation systems.
Codoop treats the vector store as one index among others, hybrid with keyword search, filtered by tenant, and re-indexed when your source documents change.
Related: /ai-services/enterprise-rag