Abstract: In this work, we proposed a web-based interface for multi-language storytelling by a social robot. Teachers of any topic can use the interface to make QTrobot approach their content in narratives. We used a chatGPT API to afford autonomously story enhancement or even complete generation. Therefore, our platform can be used in many rounds of stories, in 3 levels of story generation autonomy: (i) completely generation by chatGPT, (ii) partially generated by chatGPT (story enhancement) and (iii) no interference on the inserted story. A participatory design was performed with 5 teachers from a local school of Makers Space, where they teach children from 5 to 10 y.o. concepts of physics using LEGO kits. The majority of the participants claimed they found the solution helpful to explain the theory at the beginning of the class, since at this moment, their pupils are already excited to play with the kit and do not really pay attention to the instructions. Therefore, using the QTrobot is a potential alternative to hold their attention for longer. Furthermore, the chatGPT generation seems to solve teachers’ problem of lack of time to come out with creative stories regarding each topic. Finally, we did a study case in which we deployed our solution in two group sessions where students were learning the concepts of centripetal force and rotation by building a merry-go-round.