Account

What new tool do you want?

Text & Writing

Glossary & Terminology Consistency Checker

Keep terminology consistent: list preferred terms and the variants to avoid, then check (and optionally fix) any text against your glossary.

How to use

  1. 1Enter glossary lines as 'preferred = variant, variant' (e.g. 'email = e-mail, E-Mail').
  2. 2Paste your text to see every variant used, or tick 'Fix' to replace them.

How it's calculated

Each variant is searched as a whole word (case-insensitive). Fixing replaces it with the preferred term, keeping capitalization at the start of sentences.

Frequently asked questions

What is this useful for?

Brand names, product terms, style-guide spellings and translation glossaries — anywhere the same concept must always use the same word.

Is it case-sensitive?

Matching is case-insensitive; replacements keep an initial capital if the original had one.

Is my text uploaded?

No — checks run entirely in your browser.

Worked example: Glossary & Terminology Consistency Checker

With the glossary rule email = e-mail and website = web site, pasting "Send an e-mail from the web site. E-Mail us." reports 2 variants found: "e-mail → email (2×)" and "web site → website (1×)" — both occurrences of e-mail get counted together (including the capitalized E-Mail), because matching is case-insensitive by design, counting a style inconsistency the same way regardless of where in the sentence it happens to be capitalized. Turn on Fix with just email = e-mail against "Send an e-mail. E-mail us." and it outputs "Send an email. Email us." — notice the sentence-starting "E-mail" becomes "Email" (capital E preserved) rather than lowercase "email", because the fix logic checks whether the matched text was capitalized and re-capitalizes the replacement to match.

Why whole-word matching matters more than it sounds

The regex wrapping each variant uses Unicode-aware lookaround — (?<![\p{L}\p{N}]) and (?![\p{L}\p{N}]) — to ensure "log in" as a glossary variant doesn't accidentally match inside an unrelated longer word or get confused by surrounding punctuation, and critically, it works correctly on non-English text too since \p{L} matches any Unicode letter, not just ASCII a-z. Without this boundary check, a shorter variant could spuriously match as a substring of a longer, unrelated word — exactly the kind of false positive that would make a terminology checker useless on a large document.

Why this tool exists separately from a plain find-and-replace

A single find-and-replace only handles one term at a time and has no memory of which replacements it already made — this tool takes an entire glossary (brand names, product terms, translation-memory-style preferred terms) and checks or fixes all of them in one pass, reporting exactly which variant was found how many times before you commit to changing anything. That reporting step matters for the main real use case: a style guide reviewer or translator checking a document for drift before publishing, where seeing the list of inconsistencies first is often more valuable than blindly auto-fixing everything.

The count is variants, not hits, and parentheses are notes

A glossary line without an equals sign is ignored. A note in parentheses is stripped, so login (verb) is searched as login. Fix starts off. Variants found counts how many different variants showed up, not how many times each one showed up. The times sit in the note, such as 2 times. If nothing is off, the result says all terms match the glossary. An empty second box asks for lines like email = e-mail. When I turn Fix on, only the first letter of the replacement is capitalized, and only when the matched text started with a capital. An all-caps variant does not stay all caps. The counts are taken from the text I pasted, before the replacements are applied, so two rules can both see the original wording.

Related tools

More: Text & Writing