Because of its very conception, the G1000 in Belgium cannot be categorized as a form of constitutional deliberative democracy per se. Its grassroots origin never indeed entailed to change the constitution. Yet this chapter contends that there are some constitutional deliberative democracy features in the G1000, which paradoxically were not thought of by its citizen organizers who sought in the first phases of the G1000 to avoid any political and institutional ties. In fact, their focus was much more on a high input and throughput legitimacy, rather than a high output legitimacy. Their goal was to demonstrate that ordinary citizens, randomly selected, had a say about major social and political issues and that they were wiling and able to deliberate about them, should a design conducive to deliberation be put in place. While the G1000 scored highly on the input dimensions – the quality of representation was good and the agenda could not have been more open – and fairly highly on throughput legitimacy – with a clear script and trained moderators, but with processes of aggregation insufficiently transparent –, the outputs were in the short term very limited, which was a major source of criticism as media had fostered a climate of great expectations about the outputs. The absence of formal links to the main political actors meant that the organizers could not guarantee any formal implementation of the results. So the design characteristics that increase input legitimacy also undermine output legitimacy. But on the longer term the political uptake and the social uptake of the G1000 are increasing as, on the one hand, most of the political parties are now advocating some forms of participatory and deliberative democracy and, on the other hand, several experiences inspired by the G1000 have sparked around in Belgium and in neighboring countries. This twofold output consequence of the G1000 seems to indicate that this experience has fostered some sort of constitutional deliberative democracy broadly defined.