If the entire text is in the HTML of the paywalled article it will be extracted.
It works on a lot of paywalled articles. Not all, but many.
Bump
ff