Repositories have historically focused on a single use case, primarily the capture of traditional (text-based) open access publications, requiring separate solutions for different use cases (e.g. research data). This made capturing the wide variety of research outputs challenging at the University of Westminster, which engages in practice-based arts research (encompassing a range of disciplines, from fine and performing arts, to architecture and design and whose outputs tend to be in a non-text format) alongside traditional research. Building on a history of collaboration, Haplo and the University and its research community have built a single, open source repository meeting multiple use cases including text-based and non-text based outputs, portfolios and research data. Made possible through the flexible technical architecture of the Haplo platform, whose underlying technology is based on semantic web principles and meets COAR's vision for next-generation repositories. Improvements to the repository now enable better capture and display of research outputs across disciplines. Highlights include the development of dynamic portfolios, improved support for non-text based outputs and ongoing engagement with practice-based arts researchers to understand their needs, build workflows, review metadata and build, test and implement a transformed repository.