Many smartphone applications (apps) for mental health (MHapps) are available to the public. However, few have been the subject of a randomized controlled trial (RCT), and the change processes that are hypothesized to mediate claimed effects have not been previously studied. This RCT compared the efficacy of three publicly available MHapps to a waitlist control condition in a community sample, in which no MHapp was provided. The three MHapps included cognitive behavioural therapy (CBT) toolkit app MoodKit, mood tracking app MoodPrism, and CBT strategy app MoodMission. Participants were randomly allocated to each condition, completed a baseline assessment, downloaded their allocated MHapp, and completed a second assessment 30 days later, with n = 226 included in final analyses (81% female; M age = 34 years). Compared to the control condition, all MHapp groups experienced increases in mental wellbeing, MoodKit and MoodMission groups experienced decreases in depression, and no groups experienced effects on anxiety. Mediated regressions revealed that increasing coping self-efficacy, rather than emotional self-awareness or mental health literacy, was the underlying process contributing to effects on mental health for all three MHapps. MHapps appear to be an effective solution for improving public mental health, notably by improving users’ confidence in their ability to cope.