# Context for Translators: Use AI to summarize related modules by linking into the the code

**URL:** https://community.transifex.com/t/context-for-translators-use-ai-to-summarize-related-modules-by-linking-into-the-the-code/4324
**Category:** Ideas
**Created:** [December 14, 2024, 9:26am UTC](https://community.transifex.com/t/context-for-translators-use-ai-to-summarize-related-modules-by-linking-into-the-the-code/4324 "2024-12-14T09:26:08Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![OmarIthawi](https://yyz2.discourse-cdn.com/flex030/user_avatar/community.transifex.com/omarithawi/32/3151_2.png) [@OmarIthawi](https://community.transifex.com/u/OmarIthawi)
#### Post date: [December 14, 2024, 9:26am UTC](https://community.transifex.com/t/context-for-translators-use-ai-to-summarize-related-modules-by-linking-into-the-the-code/4324/1 "2024-12-14T09:26:08Z")

</div>

A common challenge for [our project](https://app.transifex.com/open-edx/openedx-translations/dashboard/) is when starting a new language, translators have to translate 15000 entries with very little context.

This is especially a challenge for rich languages such as Arabic or Chinese which have several closeley related syonomyns which should be used depending on the context.

The common suggestion is to provide a screenshot in Transifex and code comments e.g. `Translators: read is for reading from hard drive`. Both require a lot of effort and screenshots gets dated quickly.

I think with the current LLM technology it would be possible to provide a reasonable summary about the button or translation text by having the LLM read the code repository and provides contextual information that’s more useful that a path and a line number (`search/views.py:52`) → (`This is the placeholder for the searchbar, it appears when the text is empty`).