This paper presents the co-design and deployment of a culturally contextualized mobile interface aimed at enhancing transparency in an AI- and IoT-enabled hydroponics system within the Indigenous San community in Donkerbos, Namibia. Building on a previously implemented AI-driven hydroponic system developed in South Africa, the project introduces a web-based application tailored to local needs through Community-Based Co-Design (CBCD). The design process emphasized relational accountability, local symbolism, and transparency to foster user agency and system usability. By integrating real-time data visualization, explanatory feedback, and interactive AI assistants powered by Large Language Models (LLMs), the application enabled community members to better understand and manage the system. Our findings demonstrate that transparent AI interactions can improve user agency and promote technological acceptance. This work contributes to the field of Responsible AI by illustrating how co-design can bridge technological and cultural contexts to support sustainable agriculture and digital inclusion in marginalized rural communities.