In this paper, we propose a system that allows the user of a real-time speech synthesizer to directly manipulate the F0 contour of an utterance on-line and in real-time. The intonation is generated by the Fujisaki Model, which creates the F0 contour based on accent and phrase commands that the user needs to trigger. These input commands to the model can be generated by the user with the buttons of a wireless Mycestro 3D mouse. To evaluate the usability, a study with 16 subjects was conducted and 10 monotone sentences were manipulated in real-time using the proposed system. The results show that the majority of users were able to produce a natural sounding intonation.