With DragGAN, anyone can warp an image with precise control over pixel placement, thus manipulating the pose, shape, expression, and location of various categories such as animals, cars, people, landscapes, etc. Because these manipulations are performed on a learned GAN image generating manifold , they tend to produce realistic output even for complex scenarios such as closed content hallucinations and shape deformations that consistently follow the object's rigidity. Both qualitative and quantitative comparisons demonstrate the advantage of DragGAN over previous approaches in image processing and point tracking tasks