Overview
The Design Research Society’s library holds decades of the field’s best thinking, and almost none of it reaches the people doing the work, because it sits in dense academic abstracts. This is a searchable index of about 6,000 of those papers, each re-summarized for a working designer at an easy reading level, and tagged by how relevant it actually is to UX.
It is live, free, and runs as a single static page with client-side search. No login, no tracking.
The problem
Design research and design practice barely talk to each other. A practitioner who wants to know what is actually understood about, say, participatory design or accessibility has to wade through thousands of conference abstracts written for other academics. The knowledge exists; the on-ramp does not.
The goal was a digest a designer could browse on a coffee break and leave with three papers worth reading, not a literature review.
How it was built
Every paper’s metadata was pulled through the library’s official open interface, taking only the metadata the repository explicitly permits and never the PDFs. A local language model then summarized each one into the same shape: a one-line gist, a plain summary, why it matters for UX, key takeaways, and topic tags, with a relevance rating from high to none. Around 6,000 papers were processed on a home GPU overnight.
About 62 percent came back genuinely relevant to UX, which became the spine of the filtering: you can set the rest aside in a click.
The reading experience
The interface is built for triage. Multi-term search, filters for relevance, topic, collection, and year, and cards that open from a one-line gist to a full summary and key takeaways. A hand-picked “Start here” strip pins broadly useful papers for anyone who arrives without a query.
Because the point is reading, the typography got real attention: a capped line measure so text never runs to the edge of a wide screen, a comfortable reading size, and a contrast pass so every label and gray meets WCAG AA against the dark background.
Faithful, and honest about it
Each summary stays faithful to the paper’s own abstract rather than embellishing it. The tool is candid about edge cases too: summarizing at this scale surfaced genuine metadata errors in the source catalog itself, where a record’s title and abstract did not match.