Scope: A range of health benefits are attributed to consuming urolithin A (UA), such as improved muscle health, anti-aging activity, and neuroprotection, whereas few studies raise possible adverse effects at high doses, including genotoxicity and estrogenic effects. Therefore, understanding UA bioactivity and safety depends on its pharmacokinetics. However, there is no physiologically-based pharmacokinetic (PBPK) model available for UA, thus limiting reliable assessment of effects observed from in vitro experimentation. Methods and results: We characterizes glucuronidation rates of UA by human S9 fractions. Partitioning and other physicochemical parameters are predicted using quantitative structure–activity relationship tools. Solubility and dissolution kinetics are determined experimentally. These parameters are used to construct a PBPK model, and results are compared with data from human intervention studies. We evaluates how different supplementation scenarios may influence UA plasma and tissue concentrations. Concentrations at which either toxic or beneficial effects are previously observed in vitro appear unlikely to be achieved in vivo. Conclusion: A first PBPK model for UA is established. It enables prediction of systemic UA concentrations and is critical for extrapolating in vitro results to in vivo uses. Results support the safety of UA, but also challenge the potential for readily achieving beneficial effects by postbiotic supplementation.