In recent years, preprints—i.e., scholarly manuscripts that have not been peer reviewed or published in a journal—have emerged as a major source of research communication and a critical component of open science. However, concerns have been raised about preprints’ potential to facilitate the spread of flawed or misleading research due to the lack of quality control performed by preprint servers. Yet, there is limited knowledge of how servers currently vet incoming content and how this impacts the openness and diversity of scholarly content. In this paper, we examine preprint servers’ moderation processes, the intentions underpinning them, and their potential effects through a qualitative analysis of in-depth interviews with 14 key preprint server personnel. We find a wide range of moderation processes, which vary depending on specific server contexts and needs and are motivated by a desire to prevent the spread of misinformation and protect trust in preprints and servers. Participants repeatedly emphasized the difference between their moderation processes and peer review, but in practice often applied similar criteria for delineating scientific from unscientific content. Moreover, moderation processes often relied on trust cues, such as article formats or author affiliations, as proxies for research quality, potentially introducing similar biases as have been found in traditional journal peer review. We discuss implications for the diversity of preprint content and authors, as well as the future of preprint servers within an evolving scholarly communication ecosystem.