Embodiments are disclosed herein that relate to multi-modal interaction on a computing device comprising a multi-touch display. One disclosed embodiment comprises a method of multi-modal interaction including recognizing a hand posture of a user's first hand directed at the display and displaying a modal region based on the hand posture, wherein the modal region defines an area on the display. The method further includes receiving an input selecting a mode to be applied to the modal region, wherein the mode indicates functionalities to be associated with the modal region and defines a mapping of touch gestures to actions associated with the mode. The method further includes, while the modal region remains displayed, recognizing a touch gesture from a user's second hand directed at the display within the modal region and performing an action on the display based upon a mapping of the touch gesture.