hq

https://github.com/orf/hq

hq reads HTML and converts it into a JSON object based on a series of CSS selectors. The selectors are expressed in a similar way to JSON, but where the values are CSS selectors. For example:

{posts: .athing | [ {title: .titleline > a, url: .titleline > a | @(href)} ] }

jq, but for HTML. Similar to pup but maybe more maintained?