Je cite la doc des paramètres
char-encoding: raw, ascii, latin1, utf8 or iso2022
Determines how Tidy interprets character streams. For ascii, Tidy will
accept Latin-1 character values, but will use entities for all
characters whose value > 127. For raw, Tidy will output values above
127 without translating them into entities. For latin1 characters above
255 will be written as entities. For utf8, Tidy assumes that both input
and output is encoded as UTF-8. You can use iso2022 for files encoded
using the ISO2022 family of encodings e.g. ISO 2022-JP. The default is
ascii.