Time-to-event data refers to the observed time from a defined origin (e.g. diagnosis of a disease) until a terminating event of interest (e.g. death). Time-to-event data emerges in a range of industries and scientific disciplines, although it is particularly common in medical and pharmaceutical research. In these research fields, time-to-event data is commonly known as survival data reflecting the fact that death is an event endpoint often used in clinical studies. Analyses of survival data are widely used for decision making in clinical trials, drug development and regulatory approvals. In this talk we introduce a flexible family of Bayesian survival models that are being integrated into the rstanarm R package through the new stan_surv modelling function. The implementation uses a familiar formula syntax for specifying covariates and censoring mechanisms, based on the widely recognised survival R package. The stan_surv modelling function accommodates standard parametric (e.g. exponential, Weibull and Gompertz) survival models under either hazard or accelerated failure time formulations. Additionally, flexible parametric (cubic spline-based) hazard models are available. These allow the time-dependent baseline hazard and time-dependent effects of covariates to both be modelled using flexible smooth functions. We demonstrate the software using an example dataset. We put particular emphasis on functionality that allows practitioners to implement survival analyses as part of a robust Bayesian workflow, including prior and posterior checks and efficient leave-one-out cross-validation.