Techniques for programmatically controlling computer programs that are configured to be controlled via text-based user interfaces. The techniques include using at least one computer hardware processor to perform: accessing a string specifying content of a user interface (UI) screen of a computer program configured to be controlled via a text-based user interface; generating, based on the string, a two-dimensional (2D) representation of the UI screen; identifying a first controllable UI element in the UI screen at least in part by processing text in the 2D representation of the UI screen; and controlling the computer program by causing, via the first controllable UI element, the computer program to perform at least one action in furtherance of a task.