# Non romanized lanuages like chinese, korean japanese

**URL:** <https://discuss.elastic.co/t/non-romanized-lanuages-like-chinese-korean-japanese/376786>\
**Category:** Elasticsearch\
**Created:** [April 4, 2025, 11:18am UTC](https://discuss.elastic.co/t/non-romanized-lanuages-like-chinese-korean-japanese/376786 "2025-04-04T11:18:44Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Shabana\_Rumane](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/shabana_rumane/32/142445_2.png) [@Shabana\_Rumane](https://discuss.elastic.co/u/Shabana_Rumane)\
**Post date:** [April 4, 2025, 11:18am UTC](https://discuss.elastic.co/t/non-romanized-lanuages-like-chinese-korean-japanese/376786/1 "2025-04-04T11:18:44Z")

</div>

Along with using plugins to help tokenize documents, is there any other way to identify similar words in different script or synonyms other than specifying them in the synonyms.txt?

---

<div class="post-metadata">

**Author:** ![iulia](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/iulia/32/124658_2.png) [@iulia](https://discuss.elastic.co/u/iulia)\
**Post date:** [April 23, 2025, 8:57am UTC](https://discuss.elastic.co/t/non-romanized-lanuages-like-chinese-korean-japanese/376786/2 "2025-04-23T08:57:34Z")

</div>

Hey Shabana,

You can try vector search with our multilingual model, E5 which would understand the semantic layer of synonyms without having to build a list.

This is a great blog to get an overview of how it works (with some examples in korean as well):

> **[Multilingual vector search: Elasticsearch with E5 embedding model -...](https://www.elastic.co/search-labs/blog/multilingual-vector-search-e5-embedding-model)**
>
> Here's how multilingual vector search works and how to use Elasticsearch with the multilingual E5 embedding model, including examples.

Josh also references a previous blog of his in that article, that goes a bit more into using the plugins you described, but with a strategy of also using a language identification model in the pipeline (in case you're dealing with multiple languages in your search - you can automatically pick the best plugin):

> **[Multilingual search using language identification in Elasticsearch](https://www.elastic.co/blog/multilingual-search-using-language-identification-in-elasticsearch)**
>
> We’re pleased to announce that we are releasing language identification in Elasticsearch 7.6. Learn about use cases and strategies for searching in multilingual corpora, and how language identificatio...