Neuro-R is a real-time inference runtime library that deploys models created by Neuro-T into vision inspection automation equipment.Neuro-R integrates various APIs and provides examples for C++/C#/Python code to quickly fuse and preprocess images so that inference on a target is fast enough to meet the real-world production requirements. Neuro-R combines flowchart function with Neuro-T to enable model integration and inference output for multiple deep learning model scenarios without the need to manually design and develop the interaction logic between multiple deep learning models, which leads to realize the model integration of multiple deep learning model scenarios and inference output.