Voice training consists of personalized sessions that support transgender individuals in changing their voices (such as modifying pitch or resonance for greater alignment with gender identity). However, transgender people face widespread oppression and health disparities, which limit their ability to access quality, gender-affirming health care, including voice training. The goal of this dissertation is to create a freely available voice training application for transgender people using a participatory research methodology. I describe the formation of an online organization called Project Spectra, which designed a novel voice training application. I also present a preliminary evaluation of the app focusing on the experiences of two transgender women who used the app for about five weeks. Using quantitative and qualitative outcome measures, the study aimed to elucidate whether and how the app furthered participants' voice goals, as well as how it fit into the context of their lives. We found that participants were able to use the app in ways that were suited to their specific needs. They also shared numerous ways the app could be improved both functionally and aesthetically. This dissertation intervenes in current lines of thinking in personal health informatics (PHI). In showing how a personal health technology can be built collaboratively online through sustained, long-term engagement with a vulnerable community, I make the case for an ideological shift in the field towards social change, beyond individual change. I explore data visualizations that support user agency and self-determination. I also join ongoing conversations in human-computer interaction (HCI) critiquing the established methods of participatory design. Online collaborative making, inspired by the open source movement, could be a promising avenue for future research in the field; however, I argue that methodological shifts must be accompanied by ideological shifts (in intentions and values) and material shifts (in power and access to resources). This project would not have been possible without the collective effort and support of the Project Spectra team: transfusion, Lsomethingsomething, Coranna, Klew, Clo, Darekun, Factoid, Iris, phasedchirp, Celestmia, Luke... thank you all so much. Thanks are also due to Project Spectra's informants and collaborators: Kristine Guan, Zheanna Erose, Lillith Whitmann, and Tori Flormann. Thanks especially to all of the participants who contributed their thoughts and experiences to the research presented here.