Vector Database

A vector database is a purpose-built database designed to store, index, and search high-dimensional vector embeddings efficiently. Unlike traditional databases optimized for exact lookups, vector databases excel at finding the most similar vectors to a given query vector using distance metrics like cosine similarity or Euclidean distance.

Vector databases are a core component of RAG systems, where document chunks are converted to embeddings and stored for later retrieval. Key operations include inserting vectors with metadata, performing k-nearest neighbor (kNN) search, and combining vector search with metadata filtering. Popular implementations include FAISS (local, developed by Meta), ChromaDB (simple prototyping), Pinecone (managed production service), and Weaviate (open-source).

At scale, vector databases use approximate nearest neighbor (ANN) algorithms like HNSW or IVF to trade small amounts of accuracy for significant speed improvements, enabling sub-second search over billions of vectors.