In this article, the mobilization of a recipe is studied in the context of talk-in-interaction through the lens of a multimodal conversational approach. The analysis of a video-recorded interaction in Italian, documenting the preparation of a cake, leads me to consider the various multimodal resources that participants exploit in order to focus their attention on the written text, to select the necessary ingredients and tools, and to position themselves, as actants and social actors, towards the recommendations of the recipe, as well as towards the actions accomplished during the activity.