Mental health concerns are globally widespread, with most individuals in need never receiving care. Researchers at the intersection of computing and mental health have argued that digital mental health tools might be one avenue towards connecting people in acute distress to support. As individuals in distress interact with digital technologies, their culturally-bound expressions of distress are assessed and categorized by the platforms they interact with (including large language model chatbots, search engines, or online social media platforms), towards connecting them to support. In turn, the design of these technologies can also influence people’s illness narratives as they make meaning from distress and search for care. My work asks the question: what can marginalized people stand to gain or lose when technology (and the algorithms that underlie it) begin to mediate the relationships people have with mental health support? What do we gain or lose when the algorithm keeps the score? Towards answering this question, in this talk, I present my research work leveraging computational and qualitative approaches to understand how technology design and marginalization together influence experiences searching for care. Building on this work, I outline a broader vision for how we might integrate considerations of power, reciprocity, and justice in the digital mental health research and intervention design.