The metaverse is presented as a possible new technological iteration for the Internet. The generation of virtual universes in which the physical and the digital converge raises the question of how education will be addressed in these new systems. We find, however, pioneering exercises such as Minecraft: Education Edition. This platform is a version based on the popular sandbox video game, which was originally created by a community of teachers. The aim of this research is to analyze and describe the idiosyncratic characteristics of Minecraft as an educational platform, framing it as one of the pioneering exercises in the metaverse. To this end, we have employed a methodology that combines Multimodal Discourse Analysis with Grounded Theory and the Constant Comparative Method. As conclusions, we observe how the Minecraft Education platform reinforces from its approach pre-existing aspects from the physical world, resizing them to adapt them to its connected digital environment. These are key elements such as the identity of the participants, their ability to act within the system, creativity through lessons as a guide to the educational objectives and the community as the backbone of the process. At the same time, it presents differential components, such as the use of avatars, the transition from textual literacy to multimodal literacy, game mechanics that boost creativity or transhuman capabilities that defy physical space-time. All in all, the platform is designed for teachers, parents and managers, to whom it offers a series of benefits. Therefore, the pedagogical action will depend on their judgment and execution, especially through the elaboration of lessons and worlds, the management of the sessions and their interaction in community environments. It is their responsibility to ensure that the educational experience is truly empowering or, on the contrary, that it ends up being governed by reproductive criteria linked to symbolic violence.