The use of pedagogical agents (PAs) as learning tools within digital learning environments is rising. Previous research shows that PAs can aid learning across various domains and age groups, but their impact on learner motivation is unclear. As PAs become more integrated into learning systems, building an in-depth and theory-driven understanding of how PAs may influence learners' motivation is necessary. We used four prominent motivation theories to guide our examination of the impact of PAs on learner motivation: social cognitive theory, situated expectancy-value theory, interest theory, and self-determination theory. A total of 58 articles met our inclusion criteria. We conducted seven, three-level meta-analyses that included 28 potentially moderating variables and used correlational and hierarchical effects and robust variance estimation. Our results revealed that PAs significantly influenced self-efficacy expectations and interest but did not significantly influence other theory-driven motivational beliefs. We discussed these findings from different theoretical perspectives and provided implications for practice.