Turn a phone photo or a PDF of a document into flat, clean text and structured fields — fully locally, with no cloud OCR, no API keys, and no model training. The project detects the sheet of paper, ...