This document discusses the design of an intelligent web search agent. It notes that current search engines suffer from recall and precision problems due to the semantic gap between what users search for and the results provided. The document outlines that an intelligent search agent would need to perceive its environment through sensors, act upon that environment with effectors, and gather information without immediate presence. It describes different search algorithms the agent could use, including uninformed blind searches like breadth-first and depth-first, as well as informed heuristic searches like A* that use path cost estimates. The document provides details on how the agent would initialize its variables, perceive information by retrieving it from URLs, take action to determine if goals are met or where to search next, and