Amusing, cringey, and also painful that two of our most common formats - PDF and HTML/CSS/JS - are such a challenge to parse and display. Probably a quarter of AI compute power seems to go into understanding just those two.
I've used almost all of the ones listed including rolling my own RSS to ePub script. I've been using Readwise Reader lately which has been a great blend of vimish keys, read it later, tagging, and AI summaries.
I can confirm this, and will add: on my Pixel 6 running GrapheneOS, I have to scroll down the "6 pack" of buttons on the in-call screen to reveal the record call button. There is no UI hint that these buttons are within a scrollable element, so discoverability is poor.