Semantic search tries to find information based on meaning rather than only exact matching words.
Semantic search retrieves results based on semantic similarity or meaning.
For example, “employee vacation allowance” can match “annual leave entitlement” even when the exact words differ. Semantic search describes the search goal or behavior; vector search is one common technical method that can support it.
Think of it like asking a librarian for books about “cars” and also getting relevant material titled “automobiles.”
The librarian understands that the topics are related instead of checking only whether the same word appears on the cover.
Query: “I forgot my password”
System: finds “Resetting your account credentials”
Output: a relevant article without an exact phrase match
Query: “employee vacation allowance”
System: finds “annual leave entitlement” in the HR policy
Output: the policy section that answers the question
Semantic search: looks for related meaning.
Keyword search: primarily looks for matching words or phrases.
Semantic search: describes the goal of finding meaning-related results.
Vector search: is one technical method that can support that goal.
Semantic search: finds relevant results.
RAG: retrieves information and supplies it to a generative model before an answer is produced.
Explore next
Semantic search looks for meaning, not only matching words.
A video explainer is not available yet.