Sorry, thought it was obvious. Since it's using backend infrastructure to fetch the assets, it can crawl them as a bot in the same way that search engines do, without allowing cookies to be saved. Since scripts are often involved in the full rendering of a page, it clearly does allow for the scripts to load before snapshotting the DOM. But only the DOM and the assets and styles are preserved. Scripts are not. Most paywalls are simple scripts. If you disable JS and cookies, you'll often see the full text of an article.
I browse with scripts disabled by default and while some paywalls rely on js to block interactions after load many simply send only partial content and a login dialog.
archive.is does "something" to get the full page for sites that specifically do not send all the content to non-logged-in user agents, and it's definitely different / more complex than simply running noscript.
There are a lot of paywalls that are done server-side - for instance the Herald Sun, which is one of the biggest newspapers in Australia, does it like this. Even if you check the responses there's nothing in them but links to subscribe and a brief intro to the article.
