i see only one annotation interface(ner-character based) contains the chinese words.
So do you support asia language(lika chinese) in all annotation interface such as span,relation and denpendency parsing etc.?
Sure, Prodigy is pretty agnostic to the language and script, and if you can render the text in a browser, you can annotate it with Prodigy. spaCy comes with trained Chinese pipelines that you can use out-of-the-box: https://spacy.io/models/zh For other languages, you can start off with the blank tokenizer only (e.g. blank:ko) and collect annotations using the different workflows, e.g. spans.manual, ner.manual and so on.
(The main reason we show an example of Chinese for character-based annotation is that it's one of the cases where character-based annotation makes sense and where you might be more likely to work with a character-based model. For other languages, you typically want to work on the token level, since that's also what your model will be predicting over.)
Ok. Thanks 