[…] Here we use two survey experiments to study how persuasive messaging affects COVID-19 vaccine uptake intentions. In the first experiment, we test a large number of treatment messages. One subgroup of messages draws on the idea that mass vaccination is a collective action problem and highlighting the prosocial benefit of vaccination or the reputational costs that one might incur if one chooses not to vaccinate. Another subgroup of messages built on contemporary concerns about the pandemic, like issues of restricting personal freedom or economic security. We find that persuasive messaging that invokes prosocial vaccination and social image concerns is effective at increasing intended uptake and also the willingness to persuade others and judgments of non-vaccinators. We replicate this result on a nationally representative sample of Americans and observe that prosocial messaging is robust across subgroups, including those who are most hesitant about vaccines generally. The experiments demonstrate how persuasive messaging can induce individuals to be more likely to vaccinate and also create spillover effects to persuade others to do so as well. The first experiment in this study was registered at clinicaltrials.gov and can be found under the ID number NCT04460703. This study was registered at Open Science Framework (OSF) at: https://osf.io/qu8nb/?view_only=82f06ecad77f4e54b02e8581a65047d7.