As mentioned in our documentation, you can use Rasa with other languages than English. We know that there are bots out there in Spanish, Japanese, or Russian, for example.
In order to understand what the user is saying, you need to define a pipeline (Choosing a Pipeline). As spacy does not provide any Korean model out of the box, I would recommend that you use the
supervised_embeddings pipeline. However, not sure if you can use the
WhitespaceTokenizer for Korean. If you cannot tokenize the text using that tokenizer, you might need to write something on your own (Custom NLU Components). Hope that get’s you started!