1. Multi-year precipitation ‘legacies’ can have stronger effects on plant community composition than rainfall in the current growing season, but variation in the magnitude of these effects is not fully understood. Direct interactions between plants and animals, such as herbivory, and indirect interactions, such as ecosystem engineering (via changes in the physical environment), may influence precipitation legacies by altering mechanisms of lagged effects. However, the role of direct and indirect plant-animal interactions in determining the strength of precipitation legacies remains largely unexplored. 2. Here, we investigated effects of current growing season rainfall and precipitation legacies on grassland composition, and the influence of herbivory and ecosystem engineering interactions on these temporal dynamics. From 2009 to 2014, a period spanning high and low rainfall, we recorded plant cover in kangaroo rat exclosures and paired control plots that included both burrow and inter-burrow areas. We used linear mixed effects modeling and analysis of community dissimilarities to evaluate plant composition responses to current and previous growing season rainfall and kangaroo rat herbivory (presence of seed foraging) and ecosystem engineering (burrowing). 3. We found that community composition was more strongly affected by precipitation legacies than by current growing season rainfall. Greater precipitation in the previous growing season enhanced grass cover and reduced forb and legume cover. Kangaroo rat trophic and engineering interactions had counteracting effects on these legacies. While burrowing increased grass cover and thereby amplified the effects of previous growing season rainfall on community composition, legacies were suppressed by the presence of kangaroo rat foraging, which decreased grass cover. Further analysis revealed that kangaroo rat foraging and burrowing had conflicting effects on residual plant biomass prior to the growing season, suggesting that precipitation legacies were influenced by altered litter dynamics. 4. Synthesis. Our study demonstrates that animals can impact the strength of precipitation legacies through direct and indirect interactions with the plant species that drive lag effects. The influence of multiple types of plant-animal interactions on precipitation legacies may be important to consider for ecosystem management and when generating predictions of community composition and productivity in future ecosystems.