BACKGROUND: Observing medication ingestion through self-recorded videos (video directly observed therapy [VDOT]) has been shown to be a cost-effective alternative to in-person directly observed therapy (DOT) for monitoring adherence to treatment for tuberculosis disease. VDOT could be a useful tool to monitor short-course latent tuberculosis infection (LTBI) treatment. METHODS: We conducted a prospective randomized controlled trial comparing VDOT (intervention) and clinic-based DOT (control) among patients newly diagnosed with LTBI who agreed to a once-weekly 3-month treatment regimen of isoniazid and rifapentine. Study outcomes were treatment completion and patient satisfaction. We also assessed costs. Pre- and posttreatment interviews were conducted. RESULTS: Between March 2016 and December 2019, 130 participants were assigned to VDOT (n = 68) or DOT (n = 62). Treatment completion (73.5% vs 69.4%, P = .70) and satisfaction with treatment monitoring (92.1% vs 86.7%, P = .39) were slightly higher in the intervention group than the control group, but neither was statistically significant. VDOT cost less per patient (median, $230; range, $182-$393) vs DOT (median, $312; range, $246-$592) if participants used their own smartphone. CONCLUSIONS: While both groups reported high treatment satisfaction, VDOT was not associated with higher LTBI treatment completion. However, VDOT cost less than DOT. Volunteer bias might have reduced the observed effect since patients opposed to any treatment monitoring could have opted for alternative unobserved regimens. Given similar outcomes and lower cost, VDOT may be useful for treatment monitoring when in-person observation is prohibited or unavailable (eg, during a respiratory disease outbreak). The trial was registered at the National Institutes of Health (ClinicalTrials.gov NTC02641106). CLINICAL TRIALS REGISTRATION: ClinicalTrials.gov NTC02641106; registered 24 October 2016.