Unicode is fully supported in TouchDesigner. Unicode can be typed into parameters, DATs, Python scripts etc. Unicode encoded text files can be loaded into DATs. File paths can include any unicode character that is legal for a file path. The Text TOP and Text SOP can accept unicode characters directly. Fallback fonts will be used if a unicode character can not be displayed in the currently selected font.
Tscript does not support Unicode, so anything that goes through TScript will lose unicode characters and result in incorrect script execution.
Although many files will be loadable back into 2018.20000, files that contain any Unicode characters will not. Extra garbage characters will be introduced into strings/DATs if loading a file from 2019.10000+ that contains unicode characters into 2018.20000.
TouchDesigner saves out .txt and other file formats using UTF-8, although it should be able to load UTF-16 and UTF-32 files as well for compatibility.
Node names, channel names, parameter names, and SOP attributes are still limited to the basic alphanumeric character set.
Byte Order Marksedit
TouchDesigner looks for Byte Order Marks at the start of .txt files to determine the encoding the file uses. Any UTF8/UTF16/UTF32 .txt file must have this mark to be properly interpreted. If the BOM is missing, it will assume the file is using CP1252 encoding, which was the previous encoding older versions of TouchDesigner used.