Mobile video-based therapy, using a portable therapy device that includes a camera, a therapy application database, a processor, and a display. The camera is configured to generate images of a user, and the therapy application database is configured to store therapy applications. The processor is configured to select, from the therapy application database, a therapy application appropriate for assisting in physical or cognitive rehabilitation or therapy of the user, to invoke the therapy application, to recognize a gesture of the user from the generated images, and to control the invoked therapy application based on the recognized gesture. The display is configured to display an output of the controlled therapy application.