Vector Database

A database built to store embeddings and find the nearest ones fast, which is what makes semantic search practical at scale.

Infrastructure Intermediate 1 min read

Also known as: Vector Store

Definition

A vector database stores embeddings and answers nearest-neighbour queries over them. Exact nearest-neighbour search is linear in the number of vectors, so these systems use approximate indexes — HNSW is the common one — that trade a small amount of recall for enormous speed.

The interesting engineering is rarely the similarity search itself. It is combining vector search with ordinary metadata filters, keeping the index fresh as documents change, and chunking documents sensibly in the first place — chunking is usually where a RAG system is won or lost.

How it relates to other terms

Prerequisite Embedding A list of numbers representing a piece of data, arranged so that similar things end up close together. Part of Retrieval-Augmented Generation Fetching relevant documents at query time and putting them in the prompt, so the model answers from your data...
Referenced by
Embedding See also Retrieval-Augmented Generation Builds on this

Related terms

Inference

Running a trained model to get an output — as opposed to training, which is producing the model in the first place.

Beginner Infrastructure

Quantization

Storing model weights at lower numerical precision to cut memory and speed up inference, trading a little accuracy for a lot of resource.

Advanced Infrastructure

Was this definition helpful?

Definitions are reviewed by our editorial team. Spotted a problem? Tell us.

Keep reading

Browse the glossary Suggest a correction