Vector search
Vector search finds items by comparing their embeddings to the embedding of a query and returning the closest ones, so results match on meaning rather than on shared keywords.
Also called: semantic search, similarity search, dense retrieval
By Divya Sharma · Checked Nov 15, 2026 · How we review
Keyword search finds documents that contain the words you typed. Vector search finds documents whose embeddings sit near the query's embedding, so "car won't start" can match a page titled "Troubleshooting a dead battery". Closeness is measured with cosine similarity, dot product or Euclidean distance.
Dense retrieval
Dense retrieval is the information-retrieval name for the same idea: queries and passages are each encoded into dense vectors, and retrieval is a nearest-neighbor lookup. The term contrasts with sparse retrieval such as BM25, where most vector entries are zero and each one stands for a word. A 2020 paper on dense passage retrieval showed that this approach, trained on question and passage pairs, could beat a strong BM25 baseline for open-domain question answering.
How it runs at scale
Comparing a query to every stored vector is exact but slow. Production systems use approximate nearest neighbor indexes like HNSW, usually inside a vector database.
Limits
Vector search can miss exact matches on names, error codes and rare terms, and it always returns something even when nothing relevant exists. Many teams pair it with keyword search in hybrid search and then apply reranking.
Related terms
- Approximate nearest neighbor search
- Approximate nearest neighbor (ANN) search finds vectors close to a query without checking every stored vector. It gives up a guarantee of the exact best match in exchange for much faster search on large collections.
- Hybrid search
- Hybrid search runs a keyword search and a vector search on the same query and merges the two ranked lists, so results can match on exact words as well as on meaning.
- Embedding
- An embedding is a list of numbers (a vector) that represents a piece of text, an image or other data, produced so that items with similar meaning end up with vectors close together.
- Vector database
- A vector database stores embeddings alongside their source records and finds the stored vectors closest to a query vector, usually with an approximate index so search stays fast across millions of items.
- BM25
- BM25 is a ranking formula for keyword search. It scores each document by how often the query's words appear in it, giving more weight to rare words and less credit to long documents and repeated terms.