What is a vector database?
A vector database is a database built to store embeddings, which are numerical representations of meaning, and to quickly find the ones most similar to a query. It supports semantic search, recommendations, and retrieval-augmented generation.
Last reviewed .
How a vector database works
You convert content into embeddings and store them with a link to the original item. To search, you embed the query and ask the database for the closest vectors.
Specialized indexing makes this fast, even across very large collections.
How it differs from a regular database
A traditional database finds records by exact values or conditions, such as a matching ID or date range. A vector database finds items by similarity of meaning, so a query can match content that uses different words.
Many projects use both, with the vector database handling meaning and a regular database handling structured facts.
When you need one
You are likely to need one when building search or an AI assistant over a large body of your own content. For small collections, simpler storage with vector search features may be enough.
Common questions
What is a vector database used for?
Common uses are semantic search, recommendations, and giving language models relevant passages in retrieval-augmented generation.
What's the difference between a vector and an embedding?
A vector is a list of numbers. An embedding is a vector that represents the meaning of a piece of data, produced by an embedding model.
Do I need a vector database for RAG?
Often, but not always. It helps when you have a lot of content to search by meaning. Small projects can sometimes use simpler options.
Is a vector database a replacement for a data warehouse?
No. A data warehouse is built for analyzing structured business data, while a vector database is built for similarity search. They serve different jobs.
Want help putting this to work?
Tell us where your marketing stands today, and we'll show you what to do first.