Skip to content

docs(custom-mt): add info about splitStringsIntoChunks #557

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
andrii-bodnar merged 1 commit into from
Dec 23, 2025
Merged

docs(custom-mt): add info about splitStringsIntoChunks #557

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 27 additions & 2 deletions src/content/docs/developer/modules/ui-modules/custom-mt.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,8 @@ For Crowdin Enterprise:
"logo": "/logo.png",
"url": "/translate",
"withContext": true,
"batchSize": 10
"batchSize": 10,
"splitStringsIntoChunks": true
}
]
}
Expand Down Expand Up @@ -104,7 +105,16 @@ For Crowdin Enterprise:
</td>
</tr>
<tr>
<td><code class="whitespace-nowrap">environments</code></td>
<td><code class="whitespace-nowrap">splitStringsIntoChunks</code></td>
<td>
<p><strong>Type:</strong> <code>boolean</code></p>
<p><strong>Required:</strong> no</p>
<p><strong>Default:</strong> <code>true</code></p>
<p><strong>Description:</strong> Controls [batch processing](#batch-processing-and-file-context). If set to <code>false</code>, all strings from a file are sent to the MT engine in a single request. If set to <code>true</code> (default), Crowdin splits the strings into chunks.</p>
</td>
</tr>
<tr>
<td><code>environments</code></td>
<td>
<p><strong>Type:</strong> <code>string</code></p>
<p><strong>Allowed values:</strong> <code>crowdin</code>, <code>crowdin-enterprise</code></p>
Expand Down Expand Up @@ -171,6 +181,21 @@ https://{AppBaseUrl}/translate/?source=en&target=uk&project_id=727186&jwtToken={
</tbody>
</table>

## Batch Processing and File Context

By default, Crowdin splits source strings into smaller chunks (`splitStringsIntoChunks: true`) to optimize processing time and manage request sizes.

However, some modern MT engines and AI providers (such as XL8 or LLMs) require the full context of a file to produce high-quality translations. This is particularly important for formats like subtitles (SRT) or literary content where the translation of one sentence depends heavily on the previous one.

If you are connecting an engine that supports file-based or document-level translation:

1. Set `splitStringsIntoChunks` to `false` in your manifest.
2. Crowdin will send all strings belonging to a single file in one request.

<Aside type="caution">
When <code>splitStringsIntoChunks</code> is set to <code>false</code>, ensure your endpoint and the connected MT engine can handle large payloads, as the request body size will correspond to the size of the full source file.
</Aside>

## Handling Non-Translatable Elements by Your MT Engine

For strings containing non-translatable elements (e.g., tags, placeholders, etc.), Crowdin replaces these elements with special *notranslate* tags. This ensures that these elements remain in their original state after the string is translated by the MT engine. Crowdin uses this approach to avoid potential issues that could break exported translation files.
Expand Down