khushi-chaudhari03/Localbiz-poster ? reverse-engineered prompt

Build me a simple web app where I can upload or snap a photo of a receipt or invoice and get the important details pulled out automatically.

I want it to read the image, find the item names, quantities, prices, taxes, subtotal, and total, then show the result in a clean, easy to read layout. It should handle messy real world receipts as best it can, and if something is unclear, it should still show what it found instead of failing completely.

Please make the experience feel quick and straightforward, with a place to paste or upload an image, a button to analyze it, and a clear results section. If it helps, use the current Gemini docs online for the image extraction part and set it up so I can run it locally with my API key from an env file.