This was such an interesting project!
I had two takes on it:
First, it led me to create a twitter account which tweets everyday a Markov Chain generated text based on Brazil’s current president, Bolsonaro. I had to use python, scrape all his tweets, and create a database from where my Markov Algorithm could create something from it. Well, not my algorithm, I used some help from APIs 
That ended up being pretty fun but I wanted to do something with Javascript and something without a database, which would find random quotes to create a new quote. I ended up using node-fetch, js-Markov and quotable, three APIs which served me quite well. It’s a bit slow since for now, for every quote there is a request. So, for an algorithm like Markov chain, which needs some corpus (a list of texts, quotes) to work with, I found that it needs at least 10-20 quotes. But 5 is fast and fun enough ;D
I think that it took me about 2 days to complete both.
Here’s my repo!