wxrks automatically splits every uploaded document into segments — usually sentence-sized chunks — before translation. That segmentation drives Translation Memory matching, word counts, and how content is presented in the Editor, so getting it right matters. This article explains wxrks' default segmentation behavior, what a Segmentation Rule (.SRX) file is, and how to upload one to override the default at the Account or Organizational Unit level.
​

💡 Who is this for? This guide is for Account Admins who need to control how wxrks splits content into segments for translation within the wxrks platform.

Before a document can be translated in wxrks, it's broken into individual segments — typically one sentence or one line at a time. Segmentation happens automatically on upload, using a set of rules that decide where one segment ends and the next begins.
​

wxrks' default rules handle the vast majority of content correctly out of the box. You'd only need to change them if your content has a pattern the defaults don't anticipate — for example, a product name that ends in a period (so it gets mistaken for a sentence end), or an abbreviation specific to your industry that isn't in the default abbreviation list. Changing segmentation rules is a specialized, technical adjustment — most accounts never need to touch it.

wxrks ships with a built-in default rule set for each source language, so no setup is required for standard use. These defaults aren't visible or editable from the UI — if you need one adjusted, upload a custom .SRX file as described below, or contact [email protected] for guidance.
​

For English (US) source content, the built-in defaults include rules such as:

Every supported source language has its own default rule set tuned to that language's own punctuation and abbreviation conventions — the examples above are specific to English (US) and won't all apply the same way to every language.

SRX (Segmentation Rules eXchange) is an XML-based industry standard for defining sentence-segmentation rules, used across translation tools generally — wxrks reads real SRX 2.0 files rather than a proprietary format, so a rule file built for another CAT tool will generally work here too, and one you build for wxrks can be reused elsewhere.
​

An SRX file is a set of rules, each built from a regular-expression pattern split into a beforebreak part (what must appear right before the potential break) and an afterbreak part (what must appear right after it), plus whether that rule breaks the text there or prevents a break. A minimal custom rule that stops wxrks from splitting after "Acme Corp." might look like this:

Hand-writing SRX XML isn't required to use this feature day to day — most teams either export a rule set from another CAT tool, use an existing SRX editor to build one, or ask wxrks support or the wxrks Community for help building or reviewing a file for a specific edge case. The one hard requirement on the wxrks side is the file extension — only files ending in .srx are accepted; anything else is rejected on upload.

You can upload a custom .SRX file at two levels — the Account and the Organizational Unit — and each uploaded file is scoped to one specific source language, not to a whole account. Only one active rule file is used per source language at a time.

To change segmentation rules for your entire Account, go to Settings > Translation Settings > Segmentation. This applies to every Organizational Unit under the Account that doesn't have its own override (see precedence below).
​

Click Add new segmentation file, choose your .SRX file, select the Source Language it applies to, and turn on Set as active if you want it to take effect immediately. Click Save.

You can also configure segmentation rules for one specific Organizational Unit. An active Organizational Unit-level rule for a given source language always takes priority over an Account-level one for that same language.
​

Open the Organizational Unit, click its Segmentation tab, then Add new segmentation file the same way as at the Account level — choose the file, set the Source Language, and activate it.

You can upload more than one rule file for the same source language, as this example Organizational Unit shows — but only one file per language should be marked Active at a time. If you activate a new file for a language that already has an active one, turn the old one off to avoid ambiguity about which rules actually apply.

For a given project's source language, wxrks resolves segmentation rules in this order:
​

This is the same override pattern used by other Organizational Unit-level settings — see All about Organizational Units for how Unit-level settings generally take priority over Account-level defaults.

To remove an uploaded rule file, click the delete icon on its row in either the Account or Organizational Unit Segmentation list. Removing the active file for a language returns that language to whichever rule set is next in the precedence order above.

My upload was rejected. wxrks only accepts files with the .srx extension — any other file type is rejected at upload. Make sure you're uploading the actual SRX rule file, not the source document you want segmented differently.
​

I'm not sure my custom rules are behaving as expected. Segmentation only affects new uploads — re-uploading or re-processing a file after changing rules will apply the new segmentation; files already segmented under the old rules won't be re-split retroactively. If a segment still isn't breaking where you expect, double-check which source language you scoped the file to, and confirm it's the one marked Active for that language.