ProTox 3.0 is a webserver designed for predicting the toxicity of chemicals, incorporating molecular similarity and machine-learning models to assess 61 toxicity endpoints. The platform is freely available and accessible to all users, offering a comprehensive computational tool for toxicologists, regulatory agencies, and researchers. ProTox 3.0 models have been validated on external datasets and show strong performance, outperforming commercial software like Discovery Studio's TOPKAT and the U.S. Environmental Protection Agency's T.E.S.T. The webserver accepts 2D chemical structures as input and provides detailed toxicological profiles, including confidence scores and radar plots. It covers various toxicity categories such as acute toxicity, organ toxicity, clinical toxicity, and toxicological pathways, among others. The models are based on Random Forest and Deep neural network algorithms, using different data sampling methods to balance sensitivity and specificity. ProTox 3.0 aims to reduce the need for animal testing and support green toxicology practices by providing efficient and cost-effective toxicity predictions. The platform is user-friendly, with a self-intuitive interface, and supports multiple output formats, making it a valuable resource for researchers and regulatory bodies.ProTox 3.0 is a webserver designed for predicting the toxicity of chemicals, incorporating molecular similarity and machine-learning models to assess 61 toxicity endpoints. The platform is freely available and accessible to all users, offering a comprehensive computational tool for toxicologists, regulatory agencies, and researchers. ProTox 3.0 models have been validated on external datasets and show strong performance, outperforming commercial software like Discovery Studio's TOPKAT and the U.S. Environmental Protection Agency's T.E.S.T. The webserver accepts 2D chemical structures as input and provides detailed toxicological profiles, including confidence scores and radar plots. It covers various toxicity categories such as acute toxicity, organ toxicity, clinical toxicity, and toxicological pathways, among others. The models are based on Random Forest and Deep neural network algorithms, using different data sampling methods to balance sensitivity and specificity. ProTox 3.0 aims to reduce the need for animal testing and support green toxicology practices by providing efficient and cost-effective toxicity predictions. The platform is user-friendly, with a self-intuitive interface, and supports multiple output formats, making it a valuable resource for researchers and regulatory bodies.