The artwork is beautiful and I love the concept, but the English is atrocious! In a case like this, it might make sense to make the story portions editable or have some kind of crowd-sourced assistance.
We also have opportunity for user to feedback about text mistakes. Select wrong text, then ctrl+enter, and propose your version.
Thanks again
There's also Soylent[1] in the pipeline for which I hope they'll release an API such that you can integrate it into apps other than word. It's also based on mechanical turk but I hope they'll accept non-US customers as an intermediary.