AI has significantly altered robotic functionality, yet much research focuses on specific algorithms, overlooking the importance of the robotics system. This limits robots’ potential to improve adaptability and functionality across different scenarios. In this paper, we address a simple yet expandable robotics system based on Autonomous AI Agent, which includes three main contributions: (1) Implementing large language-based Agent methodologies to bridge theoretical and practical robotics; (2) Designing a three-process architecture for flexible, extensible systems, allowing upgrades easily and new hardware integration; (3) Incorporating a parallel scheduling system for efficient, simultaneous execution of non-conflicting tasks, improving the performance of decision-making tasks. A modular robotic system is developed based on the ROS and Fastapi, featuring expandable hardware components. Experiments examples are conducted to verify the theoretical findings of our method in practical scenarios.