The rise of deep learning technology on microcontrollers (TinyML) marks a major innovation in the deployment of artificial intelligence on the vast number of embedded platforms,which will greatly boost the development of AIoT. There are many kinds of microcontrollers,and different chip development methods and basic APIs are also different.Compared with the PC side,the development of the reasoning technology of the model is not easy,and certain embedded development experience is required.We try to combine high-level language Python with microcontrollers to provide beginner developers with a simple and reliable low-code development kit called "OpenART",which can greatly improve the development efficiency and reduce learning efforts of deploying,evaluating,and benchmarking deep learning on microcontrollers.