Vector database
A vector database stores data as high-dimensional embeddings and retrieves items by similarity, making it the search layer behind most retrieval-augmented AI systems.
Text, images, and other data can be converted into vectors — lists of numbers that place similar meanings near each other in space. A vector database indexes these vectors so you can ask for the items most similar to a query vector, which is 'search by meaning' rather than by exact keyword.
This is the retrieval half of RAG. When an AI system needs to find the passages relevant to a question, it embeds the question and asks the vector database for the nearest stored vectors. Index type, distance metric, and embedding model all affect how good those matches are.
Learn this stuff by email
New definitions, tools, and explainers when I publish them. No spam, unsubscribe anytime.