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.