[polishdude20]

Do OCR systems still detect invisible words? I would have thought by now they'd use pixel based image recognition.

[filoeleven]

I doubt they're using OCR. More likely they're using one of the many text extractors available for PDFs.

https://stackoverflow.com/questions/3650957/how-to-extract-t...

[hoerzu]

I think nowadays they directly use screen share and image recognition like https://interview.sh for example