discover-api

Intent-ranked web discovery improves RAG workflows by returning AI-scored results that match a user's search goal rather than a keyword string, reducing noise and providing more accurate retrieval for grounding and cited briefs.

You retrieve parsed page content for research by requesting optional parsed content through the discovery API, enabling the collection of content-rich, ranked web results for analysis and knowledge base building.

Yes, you can use the discovery API across REST, CLI, and SDK workflows, allowing flexible integration whether you are scripting, prototyping, or building retrieval-augmented generation into an application.

Yes, async trigger-and-poll handling is required to retrieve AI-scored relevance results, as the discovery API processes intent ranking and optional parsed content asynchronously before returning the final web search payload.

The best way to filter vendor comparison results is using parameter-aware modes and filters with specified result limits, ensuring the discovery API returns only the most relevant, AI-scored pages for your search goal.

Web search returns irrelevant results during topic coverage because standard keyword matching lacks intent ranking, a problem solved by applying AI-scored relevance filtering to match the actual research goal.