In this paper, we describe the different ingredients of the CogiText tool which can be used for building, editing, and using large corpuses for text generation under constraints à la Alamo. In Cogi-Text, AOC-posets are used as indexes that give information about the shape of the corpuses and that help to efficiently find terms for the text creation process. We give some figures about their size and the needed time for computing them and for making a specific text creation.