Abstract. In order to test the validity of ultraviolet index (UVI) satellite products and UVI model simulations for general public information, intercomparison involving three satellite instruments (SCIAMACHY, OMI and GOME-2), the Chemistry and Transport Model, Modélisation de la Chimie Atmosphérique Grande Echelle (MOCAGE), and ground-based instruments was performed in 2008 and 2009. The intercomparison highlighted a systematic high bias of ~1 UVI in the OMI clear-sky products compared to the SCIAMACHY and TUV model clear-sky products. The OMI and GOME-2 all-sky products are close to the ground-based observations with a low 6 % positive bias, comparable to the results found during the satellite validation campaigns. This result shows that OMI and GOME-2 all-sky products are well appropriate to evaluate the UV-risk on health. The study has pointed out the difficulty to take into account either in the retrieval algorithms or in the models, the large spatial and temporal cloud modification effect on UV radiation. This factor is crucial to provide good quality UV information. OMI and GOME-2 show a realistic UV variability as a function of the cloud cover. Nevertheless these satellite products do not sufficiently take into account the radiation reflected by clouds. MOCAGE numerical forecasts show good results during periods with low cloud covers, but are actually not adequate for overcast conditions; this is why Météo-France currently uses human-expertised cloudiness (rather than direct outputs from Numerical Prediction Models) together with MOCAGE clear-sky UV indices for its operational forecasts. From now on, the UV monitoring could be done using free satellite products (OMI, GOME-2) and operational forecast for general public by using modelling, as long as cloud forecasts and the parametrisation of the impact of cloudiness on UV radiation are adequate.