It uses a sort of a DSL that I came up with, and removes a lot of boilerplate.

Headless yes.

You write WSL (Web Scraping Language) to scrape 2 pages of HN:

    GOTO news.ycombinator.com | EXTRACT {.morelink, 2} {news: .storylink}

(you may certainly remove that ' , 2 ' business based on the moral system you subscribe to)