Deep Research has become its own category of software, not just a model pointed at a search box. This article compares twelve self-hosted systems and the research architectures behind them. The line that actually matters is not whether a product ships a button labelled Deep Research, but what happens after the first round of retrieval. A genuine system notices that its plan was incomplete, chases a newly discovered lead, weighs conflicting sources, and only then writes the report. Below I ...