Modern storage facilities, large or small, face numerous challenges such as misplaced packages, physical strain on workers, limited space, and environmental degradation of stored materials due to improper temperature control. These persistent issues highlight the need for a smart, automated solution capable of improving efficiency, organization, and safety within storage environments. This project introduces AISTAX — a double-sided intelligent shelving system integrated with an advanced AI-driven robotic arm designed to handle storage tasks autonomously. The system utilizes a custom-built mechanical structure with lead screw and belt mechanisms, combined with 3D-printed components for lightweight precision. A 2D vision module powered by YOLOv8n enables the robotic arm to recognize packages based on color, size, and barcode, ensuring accurate placement and retrieval. Additionally, the shelf incorporates six adaptive air-conditioning levels to maintain ideal conditions for various materials, along with a leak detection sensor that safeguards against potential storage hazards. To enhance usability, a graphical user interface (GUI) was developed, allowing operators to monitor the system, control temperature levels, and issue retrieval commands intuitively. By merging AI perception, mechanical automation, and environmental monitoring, the AISTAX system effectively addresses the major inefficiencies in traditional storage methods. The project demonstrates that affordable, scalable, and intelligent automation can be achieved using modern AI techniques and embedded systems. The outcome of this work sets a solid foundation for the future of smart warehousing, enabling safer, faster, and more sustainable storage operations. RESULTS AISTAX system successfully demonstrated high-performance automation through the integration of AI detection, precision mechanics, and environmental monitoring. Real-time testing showed the YOLOv8n model achieved average precision of 97.4% and maintained fast inference speed of 18 ms, enabling smooth object tracking at nearly 30 FPS. The dual lead-screw mechanism delivered consistent vertical and horizontal accuracy measured at ±1.1 mm and ±0.8 mm, respectively, while the adaptive gripper achieved 94% successful handling rate across various package sizes. Environmental sensors kept internal temperatures below 38°C and triggered safety responses within 0.3 seconds during leak simulations. Overall, the system operated reliably during all test conditions, demonstrating strong stability, high efficiency, and clear readiness for deployment in real world automated storage environments.