I'm working on an app that uses both PDF and web content to build a document repository that allows you to manage your reading with tags, annotations, etc: https://getpolarized.io/

The issue with memory consumption is definitely there.

But the issue for us is that if you're doing ANYTHING with 'content' you just have to use the web platform.

You need CSS, HTML and with pdf.js we have somewhat first class PDF support (though a few features are missing).

If it's a calculator app, then sure. Don't use Electron. It would be too bloated.

One feature in Polar that simply wouldn't be possible is support for web content caching.

We support downloading web pages for offline usage to allow annotation and to prevent bit rot.

It wouldn't be possible without some sort of web control but having full chromium (plus a webserver) means that you can setup a near perfect clone of the web but on the desktop.