Anonymization Module

The Anonymization Module improves the confidentiality of your documents. You define words or patterns that Wordscope hides in the source before any segment is sent to MT engines or ChatGPT. Hidden parts are automatically restored in the translation suggestions you see in the editor.

Use it for names, emails, IDs, product codes, or any client-specific strings that should not leave Wordscope in clear text.

When it runs: Anonymization is applied to all segments of the project before they are sent to MT engines and ChatGPT. Enable it with the Enabled checkbox on the Anonymization tab.

Where: Open Settings → Customization → Anonymization tab.

Only the project owner can change these parameters.

In the Suggestion panel, click the anonymization icon to view the anonymized text that was sent to the MT engines.

---

Step by step

1️⃣ Open Settings → Customization → Anonymization.

2️⃣ Tick Enabled.

3️⃣ Add simple strings under Substrings to hide (one per line) — do not end a substring with a punctuation mark.

We want to keep some product names confidential

4️⃣ For variable or complex strings, add Regex Patterns (Pattern, Flags, Description).

5️⃣ Paste sample source text in Input Text and click Anonymize to preview the Result.

6️⃣ Use Restore on the test box if you want to reverse the preview.

7️⃣ Close Settings and translate as usual — live calls to MT/ChatGPT use the masked source; suggestions show restored text.

8️⃣ Optionally save these rules in a project template for that client.

---

Substrings vs regex

Method	When to use	Example
Substrings (one per line)	Fixed names or terms that never change	Acme Corp, Jane Smith
Regex patterns	Variable IDs, emails, product references that follow a pattern	Email pattern, product SKU pattern (below)

Built-in example in the UI: anonymize email addresses with a regex such as:

[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}

Flags: gi — Description: “Anonymize email addresses”.

---

Example — variable product references

If a document contains many product codes such as PRD-12345 or PRD-99881 (a PRD- prefix followed by digits), you may want to keep them confidential, as they could reveal the identity of the company.

A sample text with product references

A single substring cannot cover every code. Use a regex:

Field	Value
Pattern	PRD-\d{5} Find a pattern starting with “PRD-” then exactly 5 digits
Flags	gi Global search, ignore case
Description	Hide product references PRD-#####

Test input:

The medical record also mentions prior exposure to PRD-33041 and PRD-91856, with no serious adverse events reported.

After Anonymize, those codes are masked in the text sent to AI engines; Wordscope restores them in the suggestions you edit.

Variants you might need:

• PRD-[A-Z0-9]{4,8} — letters and digits, length 4–8 after the prefix
• \bSKU[:\s-]?\d{6}\b — SKU 123456 / SKU:123456

---

Ask ChatGPT to write an anonymization regex

Regex (short for Regular Expressions) are fascinating because they can identify all kinds of complex patterns in text. At first, they can seem intimidating, but there is plenty of information online, and they can actually be fun to learn.

As with many things, the best way to learn is through practice. A great way to get started is to ask ChatGPT for help: simply explain what you want to find, give a few examples, and ask it to suggest a regex pattern. This makes it much easier to understand how regex works and to improve step by step.

Example prompt:

I need a JavaScript-style regex pattern (without surrounding delimiters) to identify exact product references such as: PRD-12345, PRD-99881 (always "PRD-" then exactly 5 digits).

1️⃣ Copy the Pattern and paste it as a new regex:

2️⃣ Click Add another Regex Pattern

3️⃣ Paste the regex pattern in the pattern field, and “gi” in the flags zone (g=global search, i=case insensitive) and, add a description

4️⃣ Click Anonymize to test your pattern

If you have other product references matching another pattern, simply add another regex.

Note: If you want to learn the details, simply add “Explain each element of the pattern”

---

Check in some segments

As this happens behind the scenes, you may want to check that it is working in real time:

1️⃣ Open any segment

2️⃣ hover your mouse over the confidentiality icon on top of the Suggestion panel

This icon is visible only if Anonymization is Enabled

Wordscope will show you what it sent to the MT engines in the left column of the Suggestion panel, and what it received in return.

See what Wordscope sent to and received from the MT engines.

Tips

1️⃣ Prefer fixed substrings for stable names; use regex for IDs and product codes that vary.

2️⃣ Always preview with Anonymize on real sample segments from the job.

3️⃣ Remember: masking happens before MT/ChatGPT; restoration happens for you in suggestions.

4️⃣ Save client anonymization rules in a project template so every PM on the Team reuses them.

5️⃣ Use the Suggestion panel icon to double-check what was actually sent to the engines.

Wordscope is an all-in-one CAT tool, working seamlessly on both Mac and PC, that combines multiple machine translation engines, over a dozen specialized terminology tools, and the robust assistance of ChatGPT for Translators. It provides professional translators with advanced functionalities for translation, definition, rewriting, and much more.

Try now! Log in or sign up to receive a free trial subscription!

Visit pro.wordscope.com

Trusted by thousands of professional translators in 40+ countries