With Slate.js you can extract the text from HTML or markdown.

slatejs.org

Pretty sure Pandoc has a plain text output option. There are a few 'strip markdown' packages available on NPM as well if you're using node.

github.com/michelf/php-markdown/tree/lib/Michelf have you tried looking at this?

you could also just strip the information into a plain AST structure and reuse it differently ....

so to the question at hand :) What do you actually want to achieve? just removing all markdowns? to a plain text?