Get contents
Metadata and short verbatim excerpts of up to 10 URLs.
Returns, for each URL, its title, description, publication date, source and a short excerpt from the
beginning — never the full text. With a query, the model picks the excerpts about it and scores each
page’s relevance.
Each URL has its own status: one that fails never fails the request.
Body
urlsuri[]requiredUp to 10 full URLs.
querystringWhen set, the model picks the excerpts about this query and scores each page’s
relevance.highlightsbooleanDefault:trueReturn verbatim excerpts. Defaults to
true.max_tokensinteger | nullA hard cap on model tokens for this request, minimum 2000. If it is reached, you get everything judged so far with
incomplete: true.
Response
idstringThe request id.
object"contents"Always
contents.resultsobject[]One entry per URL, in the order they were sent.
Show child attributes
urlstringThe article’s canonical URL.
statusenum<string>okorerror. A failing URL never fails the whole request.errorobject | nullWhen
statusiserror: the reason.Show child attributes
codestringA stable code.
messagestringA human-readable explanation.
titlestring | nullThe headline.
descriptionstring | nullThe description the page declares.
published_atstring | nullPublication time, ISO 8601 in UTC.
sourcestring | nullThe publication.
excerptstring | nullThe beginning of the article, short. Never the full text.
highlightsstring[]Verbatim excerpts; about
querywhen one was sent.relevancenumber | nullWith
query: the probability that the page is about it.
usageobjectWhat the request consumed.
Show child attributes
tokensintegerModel tokens. 0 when served from cache.
callsintegerModel calls.
cost_usdnumber | nullCost in US dollars, when known.
duration_msintegerTime on the server, in milliseconds.