I'm not sure if there's a misunderstanding between the question and all responses? I think what he seeks is (2)
- First creating a simple right text editor as here on Hashnode is not that hard and impossible. The problem is of course copy and paste, but instead of trying to understand every input format, try to see if it's recognizable, if not insert as plain text, and then add up known formats. It's not breaking semantics if it's made correct.
- Second I read it as he needs a visual component builder, thats for sure doable, but still not an easy task. But here the concept is to drag divs into placeholders (remember Googles start page thing 10 years ago) and click on an edit icon to edit some values in it. It wouldn't break semantics in any way but you have to make it as form inputs for every feature as the title etc.