What has been published about this?
Type a query in plain language and get web results already structured: link, title, author, date, the passages that answer it and a relevance score. Free, no signup.
Web results as data
Unlike every other Croma tool, this one does not read a public register: it is a general web search. You type a query in plain language, 3 to 300 characters, and get up to 25 results, each with its link, title, author and publication date where the page declares them, the passages of text that answer the query, and a relevance score. That format is the whole point of the endpoint. An AI agent cannot answer with sources if it first has to interpret a results page; getting the passage and the URL already separated is what turns a search into a citation. To verify a company, a person or a court case, the official register always wins. Web search covers what no register holds: news, announcements, press history and context.
Link and provenance
URL, title, author and publication date for each result, where the page declares them.
Relevant passages
The extracts from each page that answer your query, without opening the link.
Relevance and volume
A relevance score per result, ten results by default and up to 25 if you ask for them.
Common questions
Run this from your own code
One API key gives you this source and the rest of the catalog: the United States, Brazil, Mexico, Peru and Colombia, typed and ready for your agents.