← Glossary AI

Semantic Search

Searching by meaning, not exact words.

Explained simply.

Classic keyword search (what Google used to be) matches your query against exact words on the page. Semantic search matches against MEANING. Under the hood, it embeds your query and every document, then finds the ones with the closest embeddings. You can search for 'how to fix a slow laptop' and find articles titled 'speeding up your PC' or 'performance troubleshooting', even though they share zero keywords with your query.

An example.

A product search: the user types 'gift for someone who loves coffee'. Keyword search returns nothing useful. Semantic search returns espresso machines, coffee subscription boxes, and artisan mug sets - because those products are semantically close to 'coffee gift'.

Why it matters.

Semantic search is the superpower layer underneath most useful AI features you see today: smart support bots, 'similar items' recommendations, better search boxes. When paired with classic keyword search (a 'hybrid' approach), it's the gold standard.