Swales are a widespread stormwater management solution to reduce pollutant concentrations in runoff. An innovative pilot facility was constructed to evaluate the treatment efficiency of the two main types of water-quality swales, i.e. standard swales and filtering swales. Using stormwater roof runoff, without any additions or spiked with organic micropollutants, 12 runoff simulation runs mimicking frequent storm events were discharged longitudinally or laterally over the pilot swales. The performance of each swale was assessed for 4 micropollutants, i.e. zinc (Zn), glyphosate, pyrene and phenanthrene. These substances were mainly found in the dissolved phase of the stormwater runoff used to supply the pilot swales. The standard swale, constructed from a silt loam soil, partially managed stormwater runoff by infiltration. Micropollutant concentration reductions were higher in the infiltrated water (35 - 85%) than in the overflow (-13 - 66%). The filtering swale, made of a sandy central part bordered by silt loam embankments, completely managed stormwater runoff by infiltration, providing high micropollutant concentration reductions (65 - 100%). Mass load reductions were higher for the filtering swale (67 - 90% for Zn and ? 89% for organic micropollutants) than for the standard swale (33 - 73% for Zn, 19 - 67% for glyphosate and ? 50% for both pyrene and phenanthrene). For both swales, lateral inflow was often associated with significantly higher concentration and mass reductions than longitudinal inflow. Consequently, when designing swales for the treatment of micropollutants, practitioners should preferentially promote filtering swales and installations providing lateral diffuse inflow over the facility.