Neuro-T is an all-in-one platform for deep learning visual inspection projects, which can be used for a series of tasks such as project planning-->image preprocessing-->image labeling-->model training-->model evaluation-->model training-->model evaluation, provides "automatic deep learning" solution for the traditional deep learning visual inspection program. The platform provides convenient tools and a friendly GUI, integrates automatic deep learning algorithms, combined with intelligent labeling functions, one-click generation of high-performance visual inspection models, without the need for professional knowledge in the field of AI to create deep learning visual inspection models. The platform includes 8 types of deep learning models such as object detection, classification, segmentation, etc., and solves the problem of defect detection in data-lacking scenarios by generating artificial defect pictures with GAN models and unsupervised learning models.