Token Efficient Page Content Retrieval With Highlights

via “full-page-content-retrieval-with-selective-highlighting”

Neural search API — meaning-based search, full content retrieval, similarity search for AI agents.

Unique: Integrates full-page content retrieval with query-aware highlighting to reduce token usage by ~90% (per marketing claims). Highlights are computed server-side based on relevance, eliminating need for client-side processing. Supports multiple content formats (text, HTML, markdown) in single API call.

vs others: More efficient than fetching raw URLs + client-side highlighting because relevance scoring is done server-side; reduces token usage compared to passing full pages to LLMs, lowering inference costs by ~50% (per marketing claims).