The latest available version is 2.59. See the downloads page for download and installation instructions.
Bayz can be used in two ways: on a unix/linux/windows prompt, or from R (available for Mac and Windows). Use from R is limited to some standard models and is now mostly suited for using smaller data sets. See the Bayz in R section. Use from a prompt is more versatile but requires writing a bayz script, examples are in the Bayz2.5 Examples section.
Running bayz is a two-step procedure: first the main bayz program runs the Monte Carlo Markov Chain and produces a file with the posterior samples from the specified model. After that, a number of output analyses and summaries can be made, ranging from simply obtaining the posterior mean and SD of model parameters to more advanced output diagnosis such as computation of posterior distributions of functions, retrieval of the Deviance Information Criterion, etc. To retrieve parameter estimates and all other output summaries see the Output Summaries section.