A CNN model for image classification was trained on concrete images to identify whether a concrete specimen is cracked or not. The focus was mainly on the development of the backend (FastAPI) and frontend (Streamlit) systems, with less attention given to the CNN algorithm itself.