AntAir is a dataset of gridded air temperatures in 1km spatial and daily temporal resolution currently available for the years 2003-2016. AntAir was created by modelling daily air temperature from MODIS land surface temperature using machine learning algorithms. Data from 70 weather stations was used as a reference. The dataset covers the entire continent of Antarctica in 1km spatial resolution for the period 2003 to 2016. Each year consists of 365 raster layers that represent the predicted air temperature for the respective day. The raster layers are distributed using the GeoTIFF format and the Antarctic Polar Stereographic projection (EPSG 3031). The unit of the values is degree Celsius * 10. Clouded areas where no air temperature predictions could be made for are marked as no data.