AI Glossary
Semantic search
semantic search, meaning-based search
Semantic search matches documents to a query by meaning rather than by exact words — it uses embeddings and a vector database, so it finds relevant content even when the wording differs.
- It matches by meaning, not by exact words.
- It works on embeddings and vector similarity.
- It finds relevant documents even when they don't contain the query's words.
Search for "EV", and semantic search will also return a document about an "electric car" — because it compares meaning, not exact words. Classic keyword search would fail here, when the question and the document use different phrasing.
Mechanically, the query and the documents are turned into embeddings, and the system returns the passages whose vectors are closest in the vector database. This is the core retrieval step in RAG systems, where the retrieved passages become context for the model.
Related terms
Related articles