Facial composites have been used by law enforcement agencies to help identify criminals. Initially, these facial composites were sketched by forensic specialists. Later, first software systems emerged to perform this process. But on the one hand, these systems require expert knowledge and, on the other hand, composites are created by focusing on individual parts of the face and then only combining them together in the end. Zaltron et al. presented CG-GAN, a new system that generates composites based on GAN models and in an iterative manner, as recommended by the latest research. However, the system of Zaltron et al. is still difficult to use for non-experts, so we present a new, improved system. We implement a web application for the creation of facial composites. This web application is easy to use and is accessible to all, users do not have to set up the system themselves. Users can also easily edit the features of a face as we have added the ability to adjust the features with range sliders. We also made it possible for users to undo their changes. A user study was conducted to assess whether the usability improvements also contributed to better facial composites.