Semantic Search: Finding Meaning, Not Just Keywords

Why it exists

Traditional search, or lexical search, is brittle. It relies on exact keyword matches. If you search for 'fast laptop for travel,' it might miss a great review titled 'the quickest ultrabook for digital nomads' because the words don't match. Semantic search was created to bridge this gap between user intent and literal text.

The mental model

Think of it as a search engine with a thesaurus and a sense of context. Instead of just matching strings of characters, it maps your query and the documents into a shared 'meaning space.' In this space, 'fast laptop' and 'quick ultrabook' are neighbors, even though the words are different. The search is for proximity in this meaning space, not for keyword overlap.

How it works

Modern semantic search converts text—words, phrases, or entire documents—into numerical vectors called embeddings. A machine learning model, trained on vast amounts of text, learns to place items with similar meanings close to each other in a high-dimensional vector space. When you search, your query is also converted into a vector. The system then finds the document vectors in its index that are closest to your query's vector, returning those as the most relevant results.

When to use it

Use semantic search when user intent is more important than specific keywords. This is common in general web search, e-commerce ('summer dresses' should show sundresses), and question-answering systems where a user's phrasing can vary widely. It's ideal for discovery and exploration.

When not to use it

Avoid it when precision and literal matches are critical. For example, searching for a specific error code, a legal clause, or a function name in a codebase. In these cases, a traditional lexical search is more reliable and predictable because you want an exact match, not a 'conceptually similar' one.

One canonical example

A user searches an e-commerce site for 'something to keep my coffee hot on the go.' A lexical search would fail, finding no products with that exact phrase. A semantic search understands the intent. It converts the query into an embedding that represents the concept of portable, insulated beverage containers. It then finds products whose descriptions have similar embeddings, like 'Thermos' and 'insulated travel mug,' and returns those.