All guides
Concept· 2 min

What is vector search?

Vector search finds items by meaning, not exact words — by comparing numeric embeddings.

Key idea

Text (or images, audio) is embedded into a high-dimensional vector. Similar meaning ⇒ vectors close in that space.

Example

Searching 'how to reset my password' matches 'forgot my login' because their embeddings are near each other.

What to remember

  • Backbone of modern semantic search and RAG
  • Common indexes: HNSW, IVF, ScaNN
  • Combine with keyword (hybrid) search for best recall