kisakibobo/TsengScans-AI-Manga-Translation-and-Writing-Tool ? reverse-engineered prompt

Build me a web app for translating manga panels from Japanese into English. I want to be able to upload an image, have the app detect the text in the speech bubbles, translate it, and then let me edit the translation before placing it back onto a clean version of the panel.

Please include a simple and polished interface where I can see the original text and the translated text side by side. I also want a cropping mode so I can select individual speech bubbles when the text is blurry or hard to read, and some basic image cleanup to improve recognition when needed.

Use GPT 4 vision for the translation part, and make it so the app can run locally with a web frontend and a small backend API. If you need to check current docs or best practices, look them up online. Please also make the upload and translation flow feel smooth and easy to use.