<bold>Objective:</bold> Loneliness is a global health issue, but current loneliness interventions are not scalable enough to reach many who might benefit from them. Brief online interventions could greatly expand access to evidence-based loneliness interventions. <bold>Method:</bold> We conducted a preregistered three-armed trial (N = 908, ages 16-78) to compare three self-guided online interventions: a single-session intervention (SSI) for loneliness, a 3-week, three-session intervention for loneliness, and an active control supportive therapy SSI (https://ClinicalTrials.gov, ID: NCT05687162). <bold>Results:</bold> Loneliness decreased between baseline and Week 8 across all conditions (b = -5.80; d = -0.55; 95% CI [-0.62, -0.47]; p < .01), but did not decrease significantly more in those assigned to either the loneliness SSI (b = -1.27; d = -0.12; 95% CI [-0.30, 0.06]; p = .20) or the 3-week intervention (b = -0.93; d = -0.09; 95% CI [-0.27, 0.09]; p = .34) than those assigned to the control SSI. Participants found all three interventions acceptable but rated both loneliness interventions as more acceptable than the control (p < .01). Far more participants completed the 10-min control SSI (86.6%) and 20-min loneliness SSI (69.4%) than the 60-min 3-week intervention (14.9%). <bold>Conclusions:</bold> An SSI for loneliness was not significantly less effective than a longer loneliness intervention and had a much higher completion rate. Yet, against our hypotheses, neither loneliness intervention reduced loneliness more than an active control SSI did. Future work should aim to design more effective SSIs for loneliness and identify populations for which SSIs might be most helpful.