BeagleLib is a library that provides the ability to compute phylogenetic likelihoods for any model that can be specified by an instantaneous rate matrix of a continuous-time, finite-state Markov chain. Furthermore, it allows likelihoods to be computed using Graphics Processing Units (GPUs) if they are available, which can substantially speed up likelihood evaluation for models involving many states (such as amino acid or codon models). Using BeagleLib allows us to bypass writing our own C++ code for computing likelihoods.