Billions of people worldwide lack access to safe drinking water. Drinking water systems for very small communities, which usually have fewer resources than larger cities, face several challenges in providing safe water robustly. These challenges include the cost of infrastructure, the cost of water quality testing per regulations, reliance on electrical grid/energy consumption, repair issues with highly mechanized systems, and the prohibitive cost of a full-time operator to run the facility. Communities that rely on surface waters such as streams or lakes also face challenges of poor water quality during storm events – with spikes in turbidity and fecal indicators following heavy rains. Simultaneously, storms often disrupt the electrical grid. Technology and designs from the AguaClaraCornell team now provide safe drinking water to >100,000 people in Central America and India living in communities between ~1.5-15k persons. This technology holds great promise for the 139,000 small communities throughout the US but it needs to be proven to meet the world’s most stringent standards for turbidity and fecal indicators– those of the US EPA. Additionally, there are very small communities (<1000 people) that are underserved the world over who have few affordable solutions to treat raw water effectively before distribution, so even smaller plant designs are needed. In this project we performed pilot testing of a 0.5 L/s (~11,000 gallons/day) surface water treatment system (the Plantita) to demonstrate consistent production of USEPA quality drinking water. The system includes coagulant addition, flocculation, clarification, and filtration, all performed with only gravity-based energy. To enable local construction and repair, the Plantita is constructed from parts generally found at common hardware stores. For our pilot testing, we installed in-line turbidimeters for continuous turbidity readings for raw, clarified, and filtered waters as well as the “Settled” water in a tube settler following the flocculator –as a proxy for non-settlable particles. In testing spanning months, the Plantita produced filtered water that met EPA limits (NTU<0.3) even with raw water (creek water) turbidities up to 130 NTU during storm events in the local watershed. By using the information from manual coagulant dosing experiments (n = 13 day-long experiments) we were able to parameterize a model for automatic dosing of coagulant using sensors for raw water and “settled” from the tube settler device. The performance of the Plantita was also strong upon automation (n = 11 dates). Such automation holds great promise broadly for small and very small communities that cannot afford a full-time operator on-site. In subsets of the manual and automatic dosing experiments, we examined biological water quality. We performed culture-based IDEXX tests for E. coli and total coliforms and, for qPCR-based assays, we first concentrated biomass from 10-L samples using tangential flow ultrafiltration, then extracted nucleic acids and performed qPCR assays for biomarker genes from viral and bacterial fecal indicators/pathogens. For qPCR targets, only Enterococcus (EPA qPCR Method 1609) and cRaSSphage (human fecal indicator virus) were found in raw water at any point in our sampling dates but were never quantifiable in the clarified or filtered water. In summary, the Plantita has demonstrated its utility to provide USEPA quality drinking water for very small communities worldwide.