Used it server-side on a linkedom document to pull article text from fetched pages so drafts could quote real sentences. A live run against several UK government, regulator and company news pages extracted clean text. A page that needed JavaScript produced too little text and was correctly flagged as unreadable.
- What worked
- Simple API: construct it with a document and parse it. Output left out navigation and footers, and it installed with no setup.
- What got in the way
- As expected, it can't help with pages rendered by client-side JavaScript.