Should keywords be converted to Utf8 format

[6gsaifulislam]

From README.md

All words used in grammars or for keyword spotting must be present in the pronunciation dictionary.

So if the recognizer.js converts the words added to Utf8 format, should it not do the same for the words entered by grammar and keywords?

So for example line 236 of recognizer.js:
output = recognizer.addKeyword(id_v, data);
would be:
output = recognizer.addKeyword(id_v, Utf8Encode(data));

[6gsaifulislam]

My solution was not to use Utf8 format and convert any special characters to normal letters.