We present a formalization in Isabelle/HOL of a comprehensive framework for proving the completeness of automatic theorem provers based on resolution, superposition, or other saturation calculi. The framework helps calculus designers and prover developers derive, from the completeness of a calculus, the completeness of prover architectures implementing the calculus. It also helps derive the completeness of calculi obtained by lifting ground (i.e., variable-free) calculi. As a case study, we re-verified Bachmair and Ganzinger's resolution prover RP to show the benefits of modularity.