Vector Database
TL;DR
A database optimized for fast similarity search.
What does this mean?
A vector database stores embeddings and enables extremely fast similarity searches. It is the central storage system for RAG applications and semantic search.
How it works
Texts are stored as embedding vectors. For a search query, the database calculates which stored vectors are most similar to the query — in milliseconds.
Example
Pinecone, Weaviate, or Qdrant as the vector database for an enterprise chatbot that can search thousands of internal documents.
Why it matters
Without a vector database, RAG wouldn’t be possible. They’re the infrastructure behind every AI that works with proprietary data.
Related terms