This lecture explains how Gibbs samplers can be automatically constructed for Bayesian models with a structure represented as a directed acyclic graph (DAG), and introduces the JAGS package, an implementation of this approach.