Safety signals observed in early-phase clinical trials can have profound implications for the development of new medical products. It is difficult to interpret the occurrence of safety signals in small clinical trials, where formal inferential procedures may be impractical. We advocate wider use of simple Bayesian methods to explore and understand safety signals from small clinical trials. We describe a new web-based application, BayesWeb.com, that makes such exploratory Bayesian methods accessible to subject-matter experts. We illustrate the use of this software with an example from a hypothetical phase I trial for a Factor VIII product for patients with hemophilia A.