Within the field of open access (OA) publishing, community-led publishing projects are experimenting increasingly with new forms of collaboration and organisation. They do so by focusing on setting up horizontal alliances between independent projects within a certain sector (e.g., scholar-led presses), or vertically across sectors with other not-for-profit organisations (e.g., through collaborations with libraries, universities, and funders), in order to create multi-stakeholder ecologies within scholarly publishing. Yet at the same time, imaginaries for future modes of OA knowledge production are still controlled through demands for 'scalability' and 'sustainability', which are both seen as preconditions for scholarly communication models and practices to succeed and to be efficient. But they are also prerequisites to receive funding for publishing projects or infrastructure development.The scalability of open models is perceived as essential to compete in a landscape dominated by a handful of major corporate players.Drawing on our work with the Radical Open Access Collective, the ScholarLed consortium, and the Community-led Open Publishing Infrastructures for Monographs (COPIM) project, this article outlines an alternative organisational principle for governing community-led publishing projects based on mutual reliance, care, and other forms of commoning. Termed 'scaling small', this principle eschews standard approaches to organisational growth that tend to flatten community diversity through economies of scale. Instead, it puts forward the idea that scale can be nurtured through intentional collaborations between community-driven projects that promote a bibliodiverse ecosystem while providing resilience through resource sharing and other kinds of collaboration. Following Anna Tsing's recommendations to keep in mind how reimagining our knowledge practices requires we pay particular attention to articulations between the scalable and the nonscalable (Tsing, 2012), what is needed to enable this is, first and foremost, a rethinking of existing systems and infrastructures and how they currently function -systems that have historically developed and been continuously remade to encourage further scalability. We further explore the possibilities of scaling small with particular reference to Anna Tsing's work on the 'latent commons' and Massimo De Angelis' discussion of 'boundary commoning', examining how these concepts are on display within the Radical Open Access Collective, ScholarLed and the COPIM project. As we will argue, reimagining the relations within publishing beyond a mere calcula-