Identify text-free PDFs for batch OCR?

A hail mary question.
I've been looking for this feature in any software for years without success.

I'm not sure it's possible. There was some discussion earlier in the year:

I've taken a look at the link you provided.
So, currently, the answer is no.