Current Limitations

translate-kit is still in beta. It handles around 95% of a typical codebase's translatable content — that's already a lot of tedious work off your plate, but the remaining 5% still needs manual supervision.

Patterns Not Yet Supported

The scanner extracts strings from JSX text, attributes, expression containers, object properties, template literals, and conditional expressions. However, some patterns are beyond its current reach:

• Strings in variables and constants — const title = "Welcome" won't be detected unless it's used directly in JSX
• Non-JSX files — strings in plain .ts files (API responses, error messages, etc.) aren't scanned

For these cases, add the keys manually to your source locale JSON and translate-kit will translate them along with everything else.

Framework Support

• Currently supported: next-intl + Next.js
• Other i18n libraries and frameworks are not yet supported, though the core pipeline (scan, codegen, translate) is framework-agnostic by design

AI Limitations

• Plurals and gender — basic handling only. Complex ICU MessageFormat patterns need manual authoring
• Context boundaries — the AI receives project context and namespace information, but very nuanced domain-specific terms may need glossary entries
• Large batches — very large codebases benefit from adjusting batchSize and concurrency settings

Roadmap

• Close the remaining gap — support strings in variables and non-JSX files
• Support other i18n runtimes and frameworks beyond next-intl
• Translation memory across projects
• Better handling of plurals and gender-specific translations
• Per-route lazy loading of namespaces