Medication errors remain at unacceptable levels and overload the health systems. Standardized prescriptions generated by digital systems, concepts, and dosing procedures lack orientation adapted to users with unequal literacy, thus requiring strategies beyond the mere simplification of language. In this sense, the project aims to evaluate an intelligent system for generating orientation on drug prescriptions, developed by open LLMs with a dataset of texts and images validated by specialists. Unlike conventional strategies for text simplification, the proposal is to generate an intelligent response, incorporating verbal and visual language that can communicate concepts and dosing procedures effectively, safely, accessibly, and through adaptation to the cultural context, minimizing risks and supporting health system users in understanding and managing their health status. The aim is to build a B2B innovation, the target of which is administrators that are sensitive to the topic. This technology seeks to serve groups with different levels of vulnerability and with great need for support in their selfcare practices, including people with low literacy, visual and hearing impairments, and members of traditional communities.