If the key spell_checking is specified in the configuration it is taken as a directory containing *.txt files of words. Generation of the HTML webpages will include spell checking in this case.
There is one word per line. If the word is all lowercase the matching is case insensitive. If the word contains uppercase characters then the matching is case sensitive.
Any words not in the *.txt files will be listed to the file given by the key misspelt_words in the configuration.
The spell checking code makes an effort to replace some HTML entity codes with ASCII letters for instance André will be spell checked as Andre.