The recently discovered charge-flipping phasing algorithm has received growing interest in small-molecule crystallography and powder diffraction. This computational methodology differs from classical direct methods as it does not require a priori knowledge of either space-group symmetry or chemical composition and does not rely on probabilistic phase relations. Here, it is shown that the charge-flipping algorithm is capable of solving large macromolecular structures with up to approximately 6000 atoms in the asymmetric unit using suitable normalized intensity data at atomic resolution ( approximately 1.0 A). Moreover, it is demonstrated that this algorithm also provides an efficient tool for the experimental phasing of highly complex heavy-atom or anomalous scattering substructures at medium to low resolution ( approximately 2-6 A) that are frequently difficult to determine using Patterson techniques or direct methods. With the present extension to macromolecular crystallography, charge flipping has proved to be a very well performing and general phase-recovery algorithm in all fields of kinematical diffraction.