<p>Houseplants can support physical and mental well-being, yet novice caretakers often struggle to interpret a plant’s needs from delayed and ambiguous cues. To address this challenge, we present Gaia, an Arduino Uno-based social robot-planter that augments a houseplant with soil-moisture, light, and motion sensing and communicates plant state through natural-language text, an expressive digital face, and audio feedback. Rather than functioning as a separate mobile application, Gaia is designed as a co-located, anthropomorphic interface that makes plant needs legible at the point of care and encourages emotional engagement with the plant. We report two evaluation stages: an early paper-prototype usability study used to compare natural-language messages with raw sensor-value displays and refine the interface, followed by a functional-prototype user study that assessed usability and affective response. Results showed that participants more readily understood and acted on natural-language messages than on sensor readings, and that Gaia’s multimodal feedback was generally interpretable and engaging. The functional prototype received a mean System Usability Scale score of 77.2, suggesting good usability. These findings indicate that anthropomorphic, multimodal plant-state feedback can support care confidence and strengthen perceived connection in human-houseplant interaction. However, these findings are based on small, exploratory user studies and should be interpreted as indicative rather than generalizable, reflecting early-stage insights from a research-through-design process.</p>