Structure-based virtual screening represents an advanced strategy to successfully identify active compounds. To date this approach has not yet been applied to screen routinely all the huge commercial chemical libraries available. We have the protocol to screen large databases of compounds in a semi-automated fashion. Our goal is to process, for the first time, the widely used databases of commercially available compounds (billions of molecules) over many targets of interest. The collected results will allow to hunt more active commercially available compounds from the larger chemical space accessible.