Most previous water treatment trials in low-income countries have evaluated point-of-use water treatment technologies used at the household level, potentially because of a previous absence of reliable technologies appropriate to treat water at the community level in low-resource settings. A meta-analysis published in 2018 estimated that point-of-use chlorination reduces diarrhoea prevalence by 24%; however, the effect was diminished and non-significant when adjusted for bias from non-blinding. The few published blinded trials assessing the effect of water treatment interventions in low-income countries have failed to detect an effect on diarrhoea. A Cochrane review published at the start of our trial in 2015 by Clasen and others concluded that further studies were needed to evaluate the effect of chlorination at the point of collection (ie, at community level) on diarrhoea. Our results contribute new evidence to the literature in several ways. First, whereas previous blinded trials have failed to detect health effects, our trial estimated a significant reduction in child diarrhoea in a low-income setting. In addition, whereas many previous water intervention trials have focused on household- level water treatment, our findings show that a low-cost automatic point-of-collection (community-level) water treatment intervention can achieve high uptake and reduce diarrhoea in a densely populated setting. Finally, our results suggest that targeting a low chlorine residual dose (0·3–0·5 ppm) in an effort to increase taste acceptability of chlorinated water can still improve water quality and reduce the risk of diarrhoea.