Vector Database
A database optimized for storing and searching high-dimensional embedding vectors.
Definition
A vector database stores numerical representations of data (embeddings) and supports efficient similarity search using metrics like cosine similarity or L2 distance. Unlike traditional databases that look up exact matches, vector databases retrieve the semantically nearest items to a query — enabling semantic search, recommendation, and RAG.
Why it matters for AI APIs
Vector databases are the memory layer for AI applications. Semantic search, RAG, duplicate detection, and recommendation systems all rely on fast approximate nearest-neighbor search over millions of embeddings. Choosing and integrating a vector database is one of the key infrastructure decisions in AI product development.
In FastAPI AI Kit
FastAPI AI Kit supports two vector backends: pgvector (embedded in your existing Postgres) and Qdrant (a dedicated vector database). Switch between them with VECTOR_STORE env var — the RAG pipeline API is identical.