Background Depression is common and tends to be recurrent. Alternative treatments are needed that are non-stigmatising, accessible and can be prescribed by general medical practitioners. Aims To compare the effectiveness of three interventions for depression: physical exercise, internet-based cognitive behavioural therapy (ICBT) and treatment as usual (TAU). A secondary aim was to assess changes in self-rated work capacity. Method A total of 946 patients diagnosed with mild to moderate depression were recruited through primary healthcare centres across Sweden and randomly assigned to one of three 12-week interventions (trail registry: KCTR study ID: KT20110063). Patients were reassessed at 3 months (response rate 78%). Results Patients in the exercise and ICBT groups reported larger improvements in depressive symptoms compared with TAU. Work capacity improved over time in all three groups (no significant differences). Conclusions Exercise and ICBT were more effective than TAU by a general medical practitioner, and both represent promising non-stigmatising treatment alternatives for patients with mild to moderate depression. Copyright and usage (C) The Royal College of Psychiatrists 2015.