When this happens it’s often because a backend component gets rewritten and somebody decides that it’s too much work to re-implement some features for the new backend. It’s much easier to come up with a PR spiel for why removing the features is actually a good thing.
OCRmyPDF is what I use as well, had good luck with it on boardgame rulebooks that sometimes come with missing or partial embedded text. Combined with recoll and the Emacs pdf-tools mode I have it all indexed and at my fingertips.
When this happens it’s often because a backend component gets rewritten and somebody decides that it’s too much work to re-implement some features for the new backend. It’s much easier to come up with a PR spiel for why removing the features is actually a good thing.