Glossary · Page model

Structured extraction

Reading a page into typed JSON you specified — by schema, by field list, by template, or by a per-row selector map.

What is structured extraction?

Structured extraction turns a rendered page into an object matching a shape you asked for. There are four paths and only one of them uses a model: a matching per-host extraction template, a per-row selector map, or captured background XHR all resolve deterministically; a schema or field list falls back to one metered model read when nothing deterministic matches.

Why it matters

Most extraction guidance assumes a model call per page, which is the expensive default. Checking for a template first, and describing rows with selectors when a page is a listing, is usually the difference between a floor-priced call and a metered one.

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.