This work defends method infallibilism, the thesis that propositional knowledge is belief based on a infallible method. The view belongs to the family of modal analyses of knowledge that have been developed in the the last fifty years. We put these analyses in a historical perspective and characterise their form and variants rigorously. Method infallibilism is defended as the simplest modal analysis that satisfies two constraints needed to avoid the Gettier problem and what we call the necessary truth problem. We implement the view in a new formal model for knowledge and a contextualist account of knowledge attributions, though we leave open whether the latter should be endorsed.