Thanks for the invite, Sandeep! :)

Short answer — apart from the draft-js core, all functionalities of the Hashnode WYSIWYG editor, are built/maintained in-house!

After deciding to go ahead with DraftJS, we did experiment with a couple of existing solutions that were built on top of draft-js... but, we ended up building a solution from scratch! The reason for doing so was two fold:

1. We weren't quite comfortable with another layer of abstraction, and understanding that these solutions demanded; before we could integrate them into Hashnode
2. We also did not want to marry into any pre-made architectural decisions (react-rte, and draftjs-plugins) as they were not "battle-tested", so to say... since draft-js itself was a relatively new library.

Taking the above two points into consideration, most of what you see in the editor, was built from scratch in a couple of weeks.

Data conversion

What is special about Hashnode's WYSIWYG editor, is not only the editor itself; but also the data conversion that happens underneath when you save/publish the content, or edit a published article; and the neat utilities which support the whole process.

All of Hashnode's user content is stored as markdown, so we have two parsers:

1. One, for converting DraftJS' EditorState data to markdown on save. We achieve this using a forked/improved implementation of draft-js-export-markdown (github.com/sstur/draft-js-export-markdown)
2. Two, for converting markdown to EditorState on edit. Since we already use a forked implementation of marked.js ; we use that to convert it to HTML, and then we use a custom DraftJS' convertFromHTML utility (with middleware, forked from github.com/HubSpot/draft-convert) to hydrate the EditorState

The second step is also used in the editor's handlePastedText paste processor function, to handle any markdown text pasted into the editor (by converting it to appropriate EditorState representation).

Open source

If I can risk to speak on behalf of open source plans for Hashnode's editor; I'm certain that this implementation is bound to be open sourced; based on what I know about Hashnode's strong "give back to the community" culture.

Don't quote me on that, though; and I'm not sure of the ETA! 🙂