Stop maintaining i18n files by hand in 2027
Stop maintaining i18n files manually when synchronization slows your team. Compare source-driven localization with key-based workflows.
- Build
The compiler finds text
Missing - Translate
AI under a spend cap
Needs review - Review
People approve or edit
Ready - Publish
A versioned release
Live
Key takeaways
- Translation files remain useful when they fit the application and team.
- Source-driven discovery can reduce manual synchronization for supported static UI text.
- Evaluate migration costs and existing workflows before replacing a working system.
Stop maintaining i18n files by hand in 2027 when that work has become an avoidable source of mistakes. The useful question is whether your team needs to copy source text, invent keys, and synchronize files manually for every interface change.
This is an argument for examining the workflow. JSON, PO, and other translation formats remain useful. A well-maintained key-based setup can be a better choice than migrating an application to a new platform.
Why do i18n files create maintenance work?#
A developer adds a sentence, gives it an identifier, moves it into a file, changes the component, and gets the identifier into the translation process. Later, someone rewrites the sentence or removes the screen. The translation catalog needs to reflect that change.
None of these steps is automatically wrong. Explicit keys can make messages searchable and separate wording from application logic. Problems appear when the process depends on people remembering a second update, or when nobody knows whether an old key is still used.
Before replacing the format, inspect the failure. Missing extraction automation, unclear ownership, and weak review rules can each create drift. A new file extension will not solve those problems by itself.
What changes with source-driven localization?#
A source-driven workflow treats supported source text as input to discovery. Linguana's compiler identifies static UI copy during builds. You do not replace every literal heading or label with a manually named translation key.
Key-based workflow7 steps per string
- 1Invent a key
- 2Wrap text in t()
- 3Add it to en.json
- 4Upload source files
- 5Translate in the TMS
- 6Download or merge a PR
- 7Ship
Linguana3 steps per string
- 1Write the text
- 2Build: text is found and uploaded
- 3Review, then publish a release
The source sentence stays readable in the component. Translation work is associated with extracted messages and compatible context. Dynamic messages with values and rich text still require explicit handling. Application integration and target-language configuration remain real tasks.
The practical benefit is fewer manual synchronization steps for static copy. It is not an assurance that a compiler understands every runtime value or that generated translations need no review.
When should you keep translation files?#
Keep a working setup when the team already has reliable extraction, established reviewer access, and a release process people trust. Existing file-based tools can also suit applications with shared messages across platforms or strict requirements to keep the translation repository as the source of truth.
Migration has a cost. Translators lose time when identifiers, context, or approval history move poorly. Developers need to retest interpolation, plural messages, fallback, and routing. A new workflow earns its place by improving the whole process.
Avoid a bulk rewrite as your first experiment. Pilot one route or feature and compare the effort with the current system. A small measured result is more useful than an architectural preference.
What should you verify before changing the workflow?#
First, check framework support and the type of rendering your app uses. Server-rendered and client-rendered surfaces need consistent locale and catalog handling. Next, identify text that discovery cannot safely infer and plan its explicit handling.
Then check who can edit, approve, and publish. These permissions influence operational safety as much as source extraction does. Confirm the fallback when a catalog is unavailable and demonstrate how your team restores an earlier release.
Finally, estimate both migration work and ongoing usage. Compatible translation-memory reuse reduces new translation work in Linguana, but changing copy or context can require new work. Review current pricing alongside the technical pilot.
A better goal for 2027#
Choose a localization process in which a developer can change a screen, a reviewer can understand the message in context, and the team can publish deliberately. The translation-file format is one implementation choice within that process.
For an existing React application, read our localization lessons. To evaluate the source-driven approach yourself, start with the framework guide. Keep the current system until a pilot shows that the alternative solves a problem your team actually has.