A plethora of chemical substances is out there in our environment, and all living species, including us humans, are exposed to various mixtures of these. Our society is accustomed to developing, producing, using and dispersing a diverse and vast amount of chemicals with the original intention to improve our standard of living. However, many chemicals pose risks, for example of developing severe diseases, if they occur at the wrong time in the wrong place. For the majority of the chemicals these risks are not known. Chemical risk assessment and subsequent regulation of use requires efficient and systematic strategies, which are not available so far. Experimental methods, even those of high-throughput, are still lab based and therefore too slow to keep up with the pace of chemical innovation.Existing computational approaches, e.g. ML based, are powerful on specific chemical classes, or sub-problems, but not applicable on a large scale. Their main limitation is the lack of applicability to chemicals outside the training data and the availability of sufficient amounts of training data. Here, we present the ready-to-use and stand-alone program deepFPlearn that predicts the association between chemical structures and effects on the gene/pathway level using deep learning. We show good performance values for our trained models, and demonstrate that our program can predict meaningful associations of chemicals and effects beyond the training range due to the application of a sophisticated feature compression approach using a deep autoencoder. Further, it can be applied to hundreds of thousands of chemicals in seconds. We provide deepFPlearn as open source and flexible tool that can be easily retrained and customized to different application settings at https://github.com/yigbt/deepFPlearn.