Custom extraction works only for English. Though a custom model will learn from the sample data you provide, custom extraction will not work properly for languages other than English. This is because all the text analysis is done in English, from the Tokenization to Entity Recognition. 

We plan to add support for other languages in the future as more of our customers and users require it.

Please contact us if you are looking to build custom extractors in other languages. 

Did this answer your question?