doi.org
From Dumping Ground to Discovery Engine: Rethinking the Institutional Repository as Connective Open Infrastructure
Institutional repositories are often treated as passive archives: places where OA materials accumulate rather than engines that actively surface, connect, and amplify open scholarship. At Carnegie Mellon University Libraries, we are rethinking this model through a suite of interoperable integrations designed to make our repository, KiltHub, a dynamic node in the broader open knowledge ecosystem.Our initiative has three interconnected workstreams: first, we have integrated KiltHub with OA Switchboard to automatically generate metadata-only records for CMU authors' OA publications supported by transformative agreements, turning publisher notifications into repository presence without burdening authors or staff or splitting impact metrics across multiple versions of a publication. Second, we are embedding repository content within external platforms: for example, OER deposited in KiltHub are surfaced directly within POEM, an IMLS-funded open educational resource platform for AI literacy, enabling discovery in the contexts where users already work. Third, we have integrated KiltHub with the Libraries' AI-powered chatbot, enabling users to discover CMU's open research through conversational search. Our broader open science discovery strategy also includes surfacing our transformative agreements, read-and-publish deals, and green and diamond OA publications through the chatbot's integration with SciFree, helping authors understand their open access options in real time while ensuring the Libraries’ efforts and investments don’t get buried in our website. Future work includes building out CHORD, a central exchange hub for verified metadata about research data, and creating a new module for the Open Science Primer on making open science infrastructure more interoperable.
