Registered Reports are a form of empirical journal article in which study proposals are peer reviewed and pre-accepted before research is undertaken. By deciding which articles are published based on the question, theory, and proposed methods, Registered Reports offer a powerful remedy for a range of reporting and publication biases. Here we reflect on the history, progress and future prospects of the Registered Reports initiative, and also offer practical guidance for authors, reviewers, and editors encountering the format for the first time. While the key ingredients of pre-study review and results-blind acceptance are far from novel – and are already adopted independently in a variety of contexts – Registered Reports are the first mechanism to combine them into a mainstream policy that has won appeal with multiple stakeholders in the research process. We review early evidence that Registered Reports are working as intended, while at the same acknowledging that they are not a universal solution for irreproducibility. We also consider how the policies and practices surrounding Registered Reports are changing, or must change in future, to address limitations and adapt to new challenges. In spite of these caveats, we conclude that Registered Reports are promoting reproducibility, transparency and self-correction across a wide range of disciplines, and may help reshape how society evaluates research and researchers.