2 ms·

I'm not disputing that they used OCR. What's wrong is that searching text in PDFs doesn't usually involve OCR.