Google developed algorithms to automatically surface content hidden behind HTML forms on the deep web. Their goal was to index this content and include it in search results. Their approach pre-computes submissions for millions of forms across hundreds of domains and languages. Key challenges included selecting appropriate input values at scale, and efficiently navigating the large search space of possible input combinations to identify the most relevant submissions. Their techniques have surfaced several million forms, driving over 1,000 queries per second to deep web content in Google search results.