Visual manipulation of a digital object such as three-dimensional digital object manipulation on a two-dimensional display surface is described that overcomes the challenges of explicit specification of axis manipulation for each of the three axes one at a time. In an example, a multipoint gesture to a digital object is received on a display surface, which generates an axis of manipulation based on a position of the multipoint gesture relative to the digital object. Then a manipulation gesture is recognized, indicative of a manipulation of the digital object relative to the axis of manipulation, and a visual manipulation of the digital object about the axis of manipulation is generated based on the manipulation gesture.