As per request, I have added an option to the n-gram generator to treat text not solely as word-based, but also as character-based/logographic text. This enables analyzing texts in, for instance, Mandarin Chinese.
Logographic n-grams in the n-gram generator
As my knowledge of character-based languages is virtually non-existent, the feature is rudimentary at this moment, although two researchers of Mandarin Chinese did test the tool and evaluate the output. With respect to Chinese, a limitation pointed out to me by Maarten Bogaards, is that the character-based script does not have spaces, and the calculator basically treats each individual character as one word, even though words can consist of multiple characters. A Chinese sample text to test with can now also be loaded, and is a sample of Mei Yuan's 隨園詩話, taken from Project Gutenberg.
One of the things I added after testing was the removal of following punctuation marks, which are different in Chinese, namely 。,、“‘’”《》…·：？！；（）and ，. You might not see the difference in all marks, but they are non-utf-8 counterparts, which, for computers at least, are a different beast. You can also enter additional characters to exclude if you so wish.