Background and Aims: Studies using smartphone apps in treatment for alcohol dependence are lacking. This study aimed to test the consumption-reducing effects of using two app-based alcohol interventions as complement to treatment as usual (TAU). Design: Three-armed, parallel, randomised controlled trial. Setting: Specialised outpatient clinic within the Stockholm Centre for Dependency Disorders, Stockholm, Sweden. Participants: 162 alcohol-dependent adults (46% female), with no social problems or mental disorders, who had no other drug use, were included. Mean age was 50 years, and the majority were highly educated, employed and had not previously received any alcohol treatment. Interventions: Participants were randomised to (1) TAU+drink-counting app (Glasklart, n = 52), (2) TAU+app-coupled breathalyser (iBAC Pro, n = 58) or (3) TAU only (n = 52). TAU included four sessions of clinician-led psychological treatment based on motivational interviewing/cognitive behavioural therapy, combined with pharmacotherapy when requested, for 12 weeks. Measurements: The primary outcome was past 4-week heavy drinking days (HDD) at 26 weeks post-randomisation, adjusted for baseline HDD, and assessed by structured interviews using Timeline Followback. Secondary outcomes included weekly consumption, consumption patterns, phosphatidylethanol and dependence at 12 and 26 weeks. Analyses compared TAU+drink-counting app and TAU+breathalyser each with TAU alone. Findings: At 26 weeks, participants with TAU+breathalyser had statistically significantly lower HDD [incidence rate ratio (IRR) = 0.67, 95% confidence interval (CI) = 0.49, 0.91] compared with those with TAU alone. There was no evidence of any effects on HDD for those with TAU+drink-counting app, compared with TAU alone (IRR = 0.94, 95% CI = 0.70, 1.25) or of any other effects. Conclusions: In Sweden, treatment as usual (TAU) for alcohol dependence combined with a smartphone application coupled with a breathalyser was more effective than TAU alone in reducing self-reported heavy drinking. There was no evidence that TAU combined with a smartphone application for drink-counting was more effective than TAU alone.