Performing an operation comprising providing, by a virtual desktop infrastructure (VDI) client, access to an application executing on a host system to a first client system via a network, receiving, from the first client system, at least one internet protocol (IP) packet specifying attributes of an input provided by an input/output (I/O) device of the client system, receiving a timestamp from a time source, and storing the attributes of the input and the timestamp as metadata of a first image corresponding to a first frame of a video output of the application.