This short course aims at global soil data users requiring detailed, comprehensive and easy to use harmonized and/or standardized soil data. SoilGrids (www.SoilGrids.org) is a system for automated soil mapping based on Machine Learning produced jointly by ISRIC — World Soil Information and collaborating organizations. SoilGrids currently serves 3D predictions of soil physical and chemical properties and soil classes (WRB soil types, USDA suborders, soil texture classes). Access to SoilGrids is achieved via REST API, Web Coverage Service (WCS), Web Map Service (WMS) and via the GSIF package for R. With its Open Data license, SoilGrids aims at becoming the OpenStreetMap, GBIF and/or OpenWeatherMap for soil data. In addition to SoilGrids, ISRIC also serves a global compilation of reference soil profile data (that were used to produce SoilGrids) via Web Feature Service (WFS). This course will provide technical details about the system design and recommend ways to access and use global soil data for various applications (such as assessment of nutrient stocks, crop modelling, hydrological modelling, land degradation modelling and spatial planning). Examples of real-world applications of SoilGrids will also be given. Printed tutorials will be provided, together with multimedia materials (video tutorials via the ISRIC youtube channel) explaining various processing steps in detail.