Gridded offshore wind datasets are increasingly demanded due to the strong development of offshore wind power generation. Offshore wind is affected by complex coastlines and coastal orography, which interact with the winds and meteorological systems by adding regional features. Thus, high-resolution wind products are usually demanded. Moreover, the characterization of extreme events also requires high temporal resolution to capture peak winds which may affect infrastructures. We present a set of high resolution (15 km, hourly) wind data sets produced using the WRF model to downscale the ERA-Interim and NCEP/NCAR reanalyses. The simulation domain covers Europe (i.e. different wind regimes). In order to follow the driving reanalysis as closely as possible, the model was restarted daily. Results show that WRF is able to produce realistic offshore wind climatologies, including wind probability distribution, annual cycle and well-known regional winds.