A computing system includes at least one video source, a virtual desktop server and a client computing device. The virtual desktop server includes a real-time media application to provide real-time communications (RTC), a native RTC engine, a geometry tracking module, and an API code redirection module to redirect intercepted APIs of the real-time media application intended for the native RTC engine based on redirection code injected into the real-time media application. The injected redirection code defines a placeholder to indicate positioning geometry of a video stream within an RTC window. The geometry tracking module detects the placeholder within the injected redirection code. The client computing device includes a display composition module to receive the video stream and the positioning geometry of the placeholder, and to overlay the video stream over the placeholder within the displayed RTC window based on the positioning geometry.