A system generates screenshots of a graphical user interface (GUI) of an application that is displayed by target devices testing the application. Each screenshot includes an image of the GUI and metadata indicative of elements of the GUI present in the image or a state of the application or target device when the image is generated. The system determines, based on the metadata of a screenshot from a first set of the screenshots and the metadata of a screenshot from a second set of the screenshots, whether the screenshots are similar and if so, the system assigns the screenshot from the second set of screenshots to a cluster that includes the screenshot from the first set of screenshots. The system outputs an indication of the cluster (e.g., a notification or graphical indication) indicative of the similarity or discrepancy between the screenshots.