Vector databases store and search embeddings.

An embedding is a list of numbers created by a machine learning model. The useful idea is simple: similar things should end up with similar lists of numbers. Text with similar meaning, images with similar content, or users with similar behavior should land near each other in the model's number space.

This enables queries that ordinary keyword search does not handle well:

A vector database does not understand meaning by itself. The embedding model turns the input into numbers. The vector database stores those numbers, builds an index over them, applies filters, and quickly returns the closest matches.

This chapter covers how embeddings and vector search work and where vector databases fit in AI systems.