open-edge-platform/geti ? reverse-engineered prompt

Build me a local computer vision app that helps me go from raw images to a trained model and live inference in one place.

I want a simple experience where I can upload images, label them, pick a task like object detection, classification, or segmentation, then train, improve, export, and run the model locally. It should work as a desktop style app or in a container, and it should also have a Python API for people who want to script the workflow.

Please make it support fast optimization and quantization for Intel hardware, with CPU support too, and keep the flow easy for non experts. I also want a clean way to choose from a set of ready made model options, train on my own data, and then test predictions right away.

If you need to check anything, look up the latest docs online and follow the repo’s existing setup and install guidance.