# Vector Database

> A vector database stores data as numerical embeddings so an AI system can find information by meaning rather than exact keywords, the retrieval engine behind most RAG systems.

_Source: https://plenaura.com/glossary/vector-database · Last updated: 2026-06-03 · Plenaura_

## Key points

- Enables semantic search: finds relevant content even when the words differ.
- Powers RAG, recommendations, and similarity search.
- Examples include pgvector, Pinecone, Weaviate, and Qdrant.

## FAQ

### Do we need a separate vector database?

Not always, for many projects an extension like pgvector on your existing Postgres is plenty. A dedicated vector database makes sense at higher scale or with specialized search needs.