The convergence of Large Language Models (LLMs) and Internet of Things (IoT) networks opens new opportunities for building intelligent, responsive, and user-friendly systems. This work presents an edge-centric framework that integrates LLMs into IoT architectures to support natural-language-based control, context-aware prompt construction, and automated device actuation. In the proposed framework, LLMs are deployed on an edge computing device connected to an IoT gateway, enabling local processing of user commands and sensor information while improving privacy and reducing dependence on cloud services. We demonstrate the feasibility of the framework through a smart-home prototype using Llama 3 and Gemma 2B models to control basic appliances through MQTT. The preliminary results highlight the trade-off between model accuracy and inference time with respect to model size. Finally, we discuss potential applications of LLM-based IoT systems and identify key challenges related to latency, reliability, resource constraints, and scalable evaluation.