Several non-cardiovascular drugs have been withdrawn from the market due to their critical side effect of inhibiting the human ether-à-go-go related gene (hERG) K+ channels, which may lead to heart arrhythmia and death. Thus, hERG safety testing is an indispensable process that is required by the US FDA. There is considerable interest in developing computational tools to filter out potential hERG blockers in early stages of drug discovery. In this work, we describe the development of a new tool for the rapid identification of potential cardiotoxic compounds by hERG inhibition. We have compiled the largest publicly available dataset of hERG binding, containing 11,958 compounds from the ChEMBL database. Once curated, this dataset contained 4,980 compounds for modeling. Several types of QSAR models have been developed and validated according to the OECD principles. The external classification accuracies discriminating blockers from non-blockers were 0.83-0.93 on external set. Model interpretation revealed several SAR rules, which can guide structural optimization of some hERG blockers into non-blockers. Virtual screening of the WDI chemical library using selected QSAR models identified 4,945 compounds as potential hERG blockers. The developed models can reliably identify blockers and non-blockers, which could be useful for the scientific community. A freely accessible web server has been developed allowing users to identify putative hERG blockers and non-blockers in chemical libraries of their interest (http://labmol.farmacia.ufg.br/predherg).