mozilla/readability is a JavaScript library that picks out only the readable main content and metadata from HTML. It ships under the name Readability.js and is a standalone version of the readability library that powers Firefox Reader View. It handles DOM documents in both browsers and Node.js, and its license is Apache-2.0.
Instead of capturing the whole page as-is, it works as a preprocessing layer for web content collection, stripping noise such as menus and ads out of the collected web pages. After installing it from npm as @mozilla/readability, calling parse() extracts the title, the content, the text and the metadata. isProbablyReaderable() gives a quick read on whether a document is worth processing, and on Node.js it can be used together with an external DOM library such as jsdom. The repository was created on February 3, 2015 and has 11,510 stars.
That said, there are no releases published on GitHub, and no published measurements comparing its extraction accuracy against other tools. The flow score the pipeline assigned it is a low 2.1 out of 10. So rather than expecting new features or performance improvements, it is closer to a way of deciding which library to standardize the main-content extraction stage of a collection pipeline on.