Search and summaries across research papers, for people who read for a living.
A researcher runs three searches, gets a wrong summary, and goes back to the library.
Insight: structured answers to act on
The hub files every answer by topic and screen, and the model names the core reasons behind them.
Gives: Pro for a month per accepted loop
Asks: Five questions after each session
Planbrand's workspace, with answers in it. Sercery's hub has the same screens, filed under Search, Summaries, Citations and Leaving.
The weekly loop
Five questions on search, summaries, citations. Every answer is scored for how specific it is; a complaint earns what a compliment earns.
- 1Which summary was wrong, and what did the paper actually say?
- 2What did you search for that it could not find?
- 3Which citation did you check by hand, and why?
- 4What did you do next with what you found?
- 5What would make you trust a summary without opening the paper?
What Sercery learns
- Where summaries go wrong
- What cannot be found
- What earns trust
Anyone who stops gets three exit questions: what made you stop, what would have kept you, what did you expect. Scored like any answer, credited, and read first by the Insights model.
This page describes Sercery's setup on Feedbank. Figures appear only where the brand has published them.