A prior distribution that, when combined with a given likelihood, yields a posterior in the same parametric family. Enables closed-form posterior updates (e.g. Beta-Binomial, Normal-Normal, Gamma-Poisson) and is the basis for analytically tractable Bayesian models.