功能

Glossary & Consistency: Keep Translations Aligned Across Projects

Glossary & Consistency: Keep Translations Aligned Across Projects

Inconsistent terminology is one of the most common translation quality issues. When "Dashboard" is translated as "Tableau de bord" in one place and "Panneau de contrôle" in another, users lose trust. When a medical term has two different translations across your patient portal, it's a safety concern.

better-i18n's glossary management lets you define approved translations for key terms and enforce them across every project and language.


How Glossary Management Works

Define Approved Terms

Create glossary entries for terms that must be translated consistently:

  • Product terminology — "Dashboard," "Workspace," "Pipeline"
  • Brand names — Terms that should never be translated
  • Industry-specific terms — Medical, legal, financial terminology
  • UI patterns — "Sign in" vs "Log in," "Cancel" vs "Dismiss"

For each term, set the approved translation in every target language. The glossary applies across all namespaces and projects in your organization.

Never-Translate Rules

Mark terms that should remain in the source language across all translations:

  • Brand names (e.g., "better-i18n," "iPhone," "Kubernetes")
  • Technical abbreviations (e.g., "API," "SDK," "CDN")
  • Proper nouns and product names
  • Code references and identifiers

When translators encounter a never-translate term, it stays as-is regardless of target language.

Consistency Enforcement

When a translator uses a term that conflicts with the glossary — even a valid synonym — they receive a notification. This catches the subtle inconsistencies that manual review often misses: two perfectly correct translations for the same concept, where consistency matters more than individual accuracy.


AI Translation + Glossary

better-i18n's AI translation (GPT-4o, Claude, Gemini, DeepL) respects your glossary automatically:

  • Approved terms are used in AI-generated translations
  • Never-translate terms remain untranslated
  • Namespace context helps the AI understand where a string appears

This means bulk AI translations maintain your terminology standards from the first draft, reducing the review burden on human translators.

Example Workflow

  1. Build your glossary — Define key terms and their approved translations
  2. AI translates a namespace — "Translate the dashboard namespace to German"
  3. AI follows glossary — "Dashboard" → "Dashboard" (never-translate), "Workspace" → "Arbeitsbereich" (approved term)
  4. Human reviews — Translator verifies accuracy with glossary already applied

Namespace-Level Consistency

Namespaces provide context that helps maintain consistency within features:

your-org/your-project
├── auth/           → "Sign in," "Password," "Two-factor"
├── dashboard/      → "Widget," "Chart," "Filter"
├── billing/        → "Invoice," "Subscription," "Plan"
└── onboarding/     → "Step," "Welcome," "Get started"

When translators work within a namespace, they see the related strings together. This natural grouping reduces inconsistency because the translator sees the full context of how terms are used within a feature.


Review Workflows

Combine glossary enforcement with review workflows for quality assurance:

  1. AI or human produces initial translation
  2. Glossary check flags any term mismatches
  3. Reviewer verifies accuracy and approves
  4. Publish deploys approved translations

Role-based access control ensures only authorized reviewers can approve translations for publication.


Audit Trail

Every glossary-related action is logged:

  • When a glossary term was created or modified
  • Who approved the change
  • Which translations were affected

This traceability is especially valuable for regulated industries where terminology consistency has compliance implications.


Managing Glossaries via MCP

Your AI assistant can help build and maintain glossaries:

"Add 'Workspace' to the glossary as a never-translate term."

"Set the approved German translation for 'Dashboard' to 'Übersicht'."

The MCP tools handle glossary operations alongside translation management, so you can build your terminology database without leaving your IDE.


Getting Started

  1. Create glossary entries in the dashboard at dash.better-i18n.com
  2. Set approved translations for each target language
  3. Mark never-translate terms for brand names and technical terms
  4. Enable AI translation — Glossary terms are respected automatically
  5. Review and publish — Human reviewers verify with glossary enforcement active

Glossary management is available on all plans.