This project leverages AI technology to perform real-time face replacement in videos and images. Users simply need to provide a single image to replace the selected face onto the target video or image, creating a lifelike deepfake effect. It is built using the Python language along with libraries such as ONNX and ffmpeg, and achieves GPU acceleration through CUDA and CoreML. The project offers a user-friendly interface that is not only easy to operate but also includes a built-in mechanism to prevent improper use, ensuring that the generated content is legal and compliant.