Vector Database
A database optimized for storing and searching high-dimensional vectors (embeddings). Powers semantic search, RAG, and recommendations.
A vector database stores and indexes vector embeddings, enabling fast similarity search. Given a query vector, it returns the most similar stored vectors (using cosine similarity, dot product, or Euclidean distance) using algorithms like HNSW or IVF.
Vector databases power semantic search (finding documents by meaning, not keywords), RAG systems (retrieving relevant context for LLMs), recommendation engines, and clustering applications.
Major vector databases in 2026: Pinecone (managed, easy), Weaviate (open source), Chroma (lightweight), Qdrant (Rust-based, fast), Milvus (large scale), and PostgreSQL with pgvector extension (for teams already on Postgres). Many traditional databases (MongoDB, Elasticsearch) have added vector search capabilities, blurring the category.