Glossary · Definition
Vector database
A vector database stores high-dimensional embedding vectors and supports fast nearest-neighbor search (find vectors similar to a query). Foundation of RAG, semantic search, and recommendation systems.
Definition
A vector database stores high-dimensional embedding vectors and supports fast nearest-neighbor search (find vectors similar to a query). Foundation of RAG, semantic search, and recommendation systems.
What it means
Two patterns: dedicated vector DBs (Pinecone, Weaviate, Qdrant, Milvus) and SQL extensions (pgvector for Postgres, sqlite-vec for SQLite). Dedicated DBs scale to billions of vectors with sub-100ms p99 query times via HNSW or IVF indexes. SQL extensions are simpler and 'good enough' for under 100M vectors.
Advertisement
Why it matters
Vector search is the bottleneck of most RAG systems. Choose wrong and your system either scales poorly or you pay for capacity you don't need. For most teams: pgvector (you already have Postgres) is the right starting point. Move to Pinecone or Qdrant when you outgrow it.
Related free tools
Frequently asked questions
pgvector vs Pinecone?
pgvector: free, just-Postgres, fine to ~100M vectors. Pinecone: managed, scales to billions, $0.40-1+ per million vectors/mo. Start with pgvector.
Open-source self-host?
Qdrant or Weaviate. Both production-grade. Qdrant tends to be faster; Weaviate has better hybrid + multi-modal support.
Related terms
- DefinitionEmbeddingsEmbeddings are dense numerical vectors that represent the meaning of text (or images, audio) in a way that semantic similarity = vector closeness. They're the foundation of RAG, semantic search, recommendation, and clustering.
- DefinitionRAG (Retrieval-Augmented Generation)RAG (Retrieval Augmented Generation) augments an LLM with documents retrieved at query time — typically from a vector database. The LLM grounds its answer in the retrieved text instead of relying purely on training data.