SoilGrids1km is a global 3D soil information system at 1 km resolution, providing spatial predictions for various soil properties and classes. It is based on global spatial prediction models derived from 110,000 soil profiles and 75 environmental covariates. The system uses automated mapping techniques, including regression and regression-kriging, to predict soil properties such as organic carbon, pH, sand, silt, clay, bulk density, and cation-exchange capacity. The predictions are validated using 5-fold cross-validation, with accuracy ranging from 23–51%. SoilGrids1km is available for download under a Creative Commons Non Commercial license. The system was developed using the Global Soil Information Facilities (GSIF), a framework for assembling and producing global soil information. SoilGrids1km uses a combination of global soil profile data and environmental covariates to generate predictions. The system includes predictions for soil organic carbon stock, depth to bedrock, and soil groups according to the World Reference Base (WRB) and USDA Soil Taxonomy. The predictions are based on a 3D regression model that accounts for spatial and vertical variations. The system has several limitations, including weak relationships between soil properties and explanatory variables due to scale mismatches, difficulty in obtaining covariates that capture soil forming factors, and low sampling density. However, the system is highly automated and flexible, allowing for increasingly accurate predictions as new data become available. SoilGrids1km is available for download via http://soilgrids.org. The system provides global soil information that can be used in various applications, including climate modeling, agriculture, and environmental monitoring. The predictions are based on a combination of soil profile data and environmental covariates, and the system is designed to be used in a wide range of applications. The system is expected to improve in accuracy as more data become available and as new technologies are developed for soil remote sensing.SoilGrids1km is a global 3D soil information system at 1 km resolution, providing spatial predictions for various soil properties and classes. It is based on global spatial prediction models derived from 110,000 soil profiles and 75 environmental covariates. The system uses automated mapping techniques, including regression and regression-kriging, to predict soil properties such as organic carbon, pH, sand, silt, clay, bulk density, and cation-exchange capacity. The predictions are validated using 5-fold cross-validation, with accuracy ranging from 23–51%. SoilGrids1km is available for download under a Creative Commons Non Commercial license. The system was developed using the Global Soil Information Facilities (GSIF), a framework for assembling and producing global soil information. SoilGrids1km uses a combination of global soil profile data and environmental covariates to generate predictions. The system includes predictions for soil organic carbon stock, depth to bedrock, and soil groups according to the World Reference Base (WRB) and USDA Soil Taxonomy. The predictions are based on a 3D regression model that accounts for spatial and vertical variations. The system has several limitations, including weak relationships between soil properties and explanatory variables due to scale mismatches, difficulty in obtaining covariates that capture soil forming factors, and low sampling density. However, the system is highly automated and flexible, allowing for increasingly accurate predictions as new data become available. SoilGrids1km is available for download via http://soilgrids.org. The system provides global soil information that can be used in various applications, including climate modeling, agriculture, and environmental monitoring. The predictions are based on a combination of soil profile data and environmental covariates, and the system is designed to be used in a wide range of applications. The system is expected to improve in accuracy as more data become available and as new technologies are developed for soil remote sensing.