GLOSSARY
Vector Database
A database that stores text (or images) as numeric embeddings and finds the most similar items in milliseconds — the memory layer behind RAG and AI search.
Instead of matching keywords, a vector database matches meaning: every document chunk is converted by an embedding model into a list of numbers capturing its semantics, and queries find the chunks whose vectors point in the most similar direction. “Car insurance claim” and “how to file after a crash” end up close together even with no shared words.
Vector databases are the retrieval half of RAG: your documents live in one, the assistant queries it before answering. Chatbase-style support bots, NotebookLM and every “chat with your docs” tool use one behind the scenes.