The Research Agent does not stop at collecting information. The gathered facts turn into ideas, strategies, development plans, and interface descriptions — documents you can hand straight to development agents and save on prototyping.
From facts to decisions
Deep research gives you the full picture of the market: competitors, prices, customer reviews, documentation. On this foundation the agent generates ideas and strategies — not invented ones, but grounded in the collected sources, with numbers and links to the originals.
How plans are structured
Plans are written as Markdown files: a main file and per-stage files (plans/NAME.md, NAME_1.md, NAME_2.md…). Every plan carries a mandatory Verification & Validation section — the criteria for checking that the result is correct. This structure reads well for humans and development agents alike.
Scenario: competitive analysis → architecture
Our typical scenario: ask it to “collect the competitors, their customers’ reviews, and the documentation” — and in 30–60 minutes you have a detailed application architecture and interfaces built on facts: who does what, what users are missing, and what they pay for. From there, implementation takes a few hours in plan-then-execute mode, with quality good enough for a prototype.
Example plans
Live examples ship with the product: research roadmaps, development plans, and a business strategy. See the Built-in Research section — the planning documents are part of the library.