In one embodiment, a server system receives, from a client device configured to remotely access a desktop hosted by the server system, user input directed to the desktop. The server system further identifies a desktop operation to be performed in response to the user input, where the identifying is performed without relying on preconfigured information that indicates what the desktop operation should be, determines when the desktop operation has completed, and adds, upon completion of the desktop operation, one or more markers to the desktop. The server system then transmits an image of the desktop that includes the one or more markers to the client device, thereby signaling the completion of the desktop operation to the client device.