No. Look at the cookbook application (http://weboob.org/applications/qcookboob). It incorporates multiple sources of recipes behind a single, consistent, tailor-made interface. This pattern repeats in the other applications. They are not taking a single website and transforming it into a native application. They are building applications to service specific domains that use the web as a data source.

They are also bringing web content into the command line. They have a large number of console applications. For example, there's a command and REPL for tracking shipments.

Putting aside the misunderstanding of what this actually is, the premise of your question seems to be that "simply replacing" one technology with another can't be "liberating." I think that's wrong in general as well for the specific example of transforming a web app to a native one (though these seem to be Python, not C++). The web has constraining properties: browsers that provide rich interaction with modern web content are resource intensive; different websites have wildly different accessibility stories (and the lack of UI and terminological consistency in content is, itself, a hurdle for some people); ads, popups, and other distractions are harmful to load times, network usage, and user experience; and on and on. There's plenty of examples of transformations from one technology to another that trade out the source's constraints for the destination's.

I think you actually misunderstood my comment. You know what are also “applications to service specific domains that uses the web as a data source”? Web pages.

Deliver the QML over the wire here and you just reinvented a kind of browser, that consumes html|json data from the web and has a different native UI library.

Web pages use a data source behind the scenes, but web pages are NOT a data source. A rendered Qt app is not a data source either. weboob builds plain-old objects (or JSON), it is a data source. The Qt applications are just an example front-end using weboob data source as a data source. Furthermore, this data source is made for aggregation/standardization, as weboob returns results in the same format whatever the site you choose, so it's still better than if each site proposed an API specific to the site itself.

If this is done half well, it will liberate the CPU cycles from tons of pointless JS.

Also, probably less reliant on analytics and third parties, unless they are baked into core APIs.

my gut says this style of application lends itself to less abuse by trackers and surveillance tech than a general purpose browser, but i could just be naive.