AR is designed to be experienced live, requiring the location and orientation of the participants head to be tracked in real time (using some combination of electro-magnetic, sonic, inertial, and video sensing). If the participant wears a camera on their head (as they must do when using a video-mixed HMD), tracking information may also be gathered by attaching fiducial marks to landmarks or objects and having the computer look for this specially designed markers in the video stream. When the computer see one of these fiducials, it can tell where the object is relative to the camera.

An AR application reacts to a participant's location and activity moment by moment, and uses the available tracking and sensor information to decide what the user sees and hears. For example, if the user is touring a university campus, the computer might draw the names of the buildings “in the air” together with arrows associating each name with the right building.

Different kinds of experiences require different kinds of content. For many experiences that might be created with DART, the system draws three-dimensional objects in the world around the user. DART also supports less common forms of content, such as video-based "actors" created by texture mapping digitized video, usually of a human actor, onto a plane in the 3D scene.

For examples of AR experiences created with DART, see Applications.