It is a simple setup. The frontend is NextJs. The backend is Python. We use an OCR framework to get the text out of the image/pdf uploaded by the user and then OpenAI to interpret the data and structure it the way we want.

Simple, right?