The AURORAE (Air qUality foRecast fOR Arctic communitiEs) service has been developed within theproject in order to produce short term forecasts of atmospheric concentrations of particulate with 10μm of equivalent diameter or less called PM10. At the daily frequency AURORAE downloads latest PM10observations from selected air quality monitoring stations located in the European Arctic and near-Arctic and PM10 and meteorological forecasts from nine CAMS (Copernicus Atmosphere MonitoringService) models over Europe. AURORAE then applies a selected AI (Artificial Intelligence) modelcombining the latest observations and CAMS model forecasts in the input and producing the outputwith PM10 forecasts up to 48 hours in advance that are more accurate than original CAMS forecasts.The forecasts are accessible through a graphical interface tailored for general users. The graphicalinterface also provides easily accessible information on historical evolution of pollution at monitoringstations. The Web access will be available from the Arctic PASSION Web site(https://arcticpassion.eu)1. Here we present a report on the description of the AI model, itsperformance, the implementation of the operational procedure and the design of the Web interface2.In the appendix we also present the AURORAE’s operational performance in year 2024 and January2025 and compare it to the operational performance of CAMS models.