A current challenge in ecology is to understand and predict how species abundance and diversity change with rapidly shifting temperature regimes. Temperature affects the local and regional dynamics underlying community structure in predictable ways, but whether this translates to predictable community-level outcomes with warming remains an open question. We address this gap with a framework that draws from metabolic theory to model temperature-dependent metacommunity processes—density-independent population growth, density-dependent biotic interactions, and dispersal—and their effects on diversity patterns. We simulated metacommunity dynamics to show how species richness, relative abundance, and beta diversity respond to up to 10°C warming, exploring effects of variation in thermal response curves for different metacommunity processes. We demonstrate how factors such as thermal safety margins and thermal asymmetries among metacommunity processes mediate the magnitude of this change. Our framework highlights the importance of incorporating dispersal and species interactions, as well as accurately identifying their thermal response curves, in predictive models of diversity change.