Make localization a content workflow, not grep-and-replace
Extract strings into locale-aware models, copy locales, translate content, and keep parity visible through validation and health checks.
Problem
Localization fails when strings stay in source files
If UI text is scattered through components, translation work becomes source editing. Contentrain moves those strings into locale-aware content files where validation and review can happen normally.
- Extract hardcoded text before translation
- Keep locale files in the repository
- Review translation diffs in Git
Modeling
Choose the right model for each language asset
Short labels often fit dictionaries, pages fit documents or collections, and global UI settings fit singletons. Contentrain lets the model reflect the content type instead of forcing every string into one table.
- Dictionaries for repeated UI labels
- Documents and collections for long-form content
- Singletons for global content surfaces
Validation
Validate locale parity and required fields
Translation quality starts with knowing what is missing. Contentrain validation detects required field issues and i18n parity problems so teams do not discover gaps in production.
- Required localized fields
- Structural validation
- Reviewable content changes
Agent assist
Let agents translate inside a governed workflow
Agent-assisted translation is useful only when the agent sees the schema, context, and vocabulary. Contentrain rules and skills provide that context before localized content is written.
- Vocabulary and context files
- Translation workflow skills
- SEO and content quality rules per locale
Team workflow
Studio gives translators and reviewers a surface
When localization involves non-developers, Studio can provide content editing, branch review, role controls, and media context without requiring translators to work directly in Git.
- Content panels for localized entries
- Reviewer roles and branch diffs
- Media and delivery context in the same workspace
Common questions
Can Contentrain manage multiple locales?
Yes. Models and content are locale-aware, and validation can catch structural and parity issues across locales.
Should we translate before or after Normalize?
For hardcoded UI text, normalize first. Once strings are content, translation can happen with better context and review.
Can AI help with translations?
Yes, but the key is governance: agents should use project context, vocabulary, schema rules, and review workflows rather than translating disconnected strings.
Next paths
Continue through the content system
Start local. Scale to Studio.
Build a governed content layer before content becomes product debt.
Developers can start with the MIT packages. Teams can move into Studio when review, roles, delivery, and licensing matter.