I suppose I can write a scraper and dump content in json format.

No need, reddit already provides their data in JSON form. Generally just append .json at the end of the URL and you get your JSON, for example

​

https://www.reddit.com/r/emacs/comments/17u00j0/extracting_forums_posts_like_reddit_discussions/

->

https://www.reddit.com/r/emacs/comments/17u00j0/extracting_forums_posts_like_reddit_discussions.json