Someday this might be a context-aware preprocessing wrapper for unidecode. As of now, the only thing it does is to call Unicode::normalize_KD before Unidecoder::decode; that makes it slightly more powerful:

irb> require "rubygems"
irb> require "unidecode"
irb> "éç™①".to_ascii
"ec"
irb> require "chou_unidecode"
irb> "éç™①".to_ascii
"ecTM1"

Just require the gem and call to_ascii in strings as usual.