lyowang/Real-ESRGAN ? reverse-engineered prompt

Build me a simple web app and HTTP API for Real ESRGAN that lets people upload a photo or short video and get back a sharper, higher resolution version.

I want it to feel easy to use, with a clean page where someone can drop in a file, choose the upscale amount, and run enhancement. It should handle common image types, videos, and also work well for anime images and anime video if the right model is selected. If possible, include a face enhancement option too, since that’s useful for portraits.

Please make the result downloadable from the browser, and also expose an endpoint so other apps can send in a file and get the processed output back. Use the existing model files in the project, and keep the app friendly for local use. If anything about the current setup is unclear, look up the latest docs online and follow the repo’s intended way of running inference.