Glossary · Page model

Extraction template

A per-host or generic-metadata recipe that reads a known page shape into structured data with no model call at all.

What is extraction template?

An extraction template encodes how a particular site — or a general convention like schema.org metadata — lays out its data, so the fields can be read deterministically. Templates are matched per host and are auto-selected unless you force one by name or disable the library, and the response says which template supplied the result.

Why it matters

A template is the cheapest possible extraction: zero model cost, no hallucination surface, and a stable result across runs. The catalog is public, so checking whether one already covers your target costs nothing and often removes the need for a schema entirely.

Related terms

See it in context

Back to the full glossary

Put your agent to work. Keep the decision.

Give your agent a real browser, authorize the accounts it may use, and keep control of every step. Repeated runs compile into skills that replay without a model.