TL;DR: Combine a web-browsing autonomous agent (one that can click, scroll, and interact) with a browser-integrated memory/summary tool (like Recall) and you get near-autonomous research: the agent finds sources and the capture tool ingests summaries a…