Parallel Web Systems: Search, Extraction, and Deep Research for Agents
Parallel Web Systems separates Search, Extract, and Task APIs into web-access layers with different latency and cost profiles, while Basis maps citations, excerpts, and confidence to output fields.