Polygraphs are junk science and do not work.

Notepad

Tesseract OCR

I am very unclear on what happens with the OCR output and how that is associated with the file and how it os searched?

Also a noob, I have just started looking at Cloudflare tunnels. Why do you need Nginx if you are using a tunnel? Can’t you map the tunnel directly to the service? I do not understand what value Nginx offers in this scenario. Maybe someone with some experience here can explain this.