Hi, I use the translation_service object from a Plone Python product to get the translaltion of texts with this method... def translate(text) ts = getattr(self, 'translation_service', None) if not ts: return text translated = ts.translate('newsletter', text) if not translated: return text else: return translated It works but returns the translated text in the original charset (ascii, latin1, ...) from the MessageCatalog and not in my plone publishing charset (utf-8). I need to know the charset used in the associated MessageCatalog object to recode that translated text into the plone publishing charset. There's certainly a way to do this because Plone recodes all that comes from a MessageCatalog through a TranslationService into the Plone publishing charset, but I'm stucked hacking TranslationService and Localizer code ! Any hint including RTFM welcome Cheers --Gilles