Abstract Solar surface irradiance (SIS) and direct (SID) irradiance as well as effective cloud albedo (CAL) climate data records (CDR) derived from the Meteosat first generation satellites (Meteosats 2 to 7, 1983–2005) are presented. The CDRs are available free of charge for all purposes from wui.cmsaf.eu at monthly, daily and hourly means at a spatial resolution of 0.03 ∘ . The processing employed a climate version of the Heliosat algorithm combined with a clear sky model using an eigenvector look-up table method. Modifications to the Heliosat method include a self-calibration algorithm as well as a running mean based clear sky retrieval algorithm. The datasets are validated using ground based observations from the Baseline Surface Radiation Network (BSRN) as a reference. The validation threshold for the mean absolute bias between satellite-derived and surface-measured radiation is given by the target accuracy for solar irradiance fields defined by the Global Climate Observing system (GCOS) and a measurement uncertainty for the surface data. The results demonstrate that the target accuracy is achieved for monthly and daily means. Furthermore, an intercomparison with similar datasets reveal a better performance and climate monitoring potential of the CM SAF SIS CDR at most BSRN sites compared to established data sets like e.g. ERA-reanalysis, GEWEX (Global Energy and Water Cycle Experiment) and ISCCP (International Satellite Cloud Climatology Project). Lastly, the realistic representation of both seasonal and inter-annual variability guarantees the applicability of the satellite-based climate data sets for climate monitoring and analysis of extremes. No trends in the normalized bias between the CM SAF and the BSRN datasets are detectable, which demonstrates the stability and homogeneity of the global and direct irradiance for the period covered by BSRN measurements. However, inconsistencies are detectable at few satellite transition dates for certain regions in earlier years.