init

The init command runs an interactive wizard to set up translate-kit in your project.

Usage

translate-kit init

What It Does

The wizard guides you through:

1. Translation mode — keys mode (t("key")) or inline mode (<T id="key">text</T>)
2. AI provider selection — choose from OpenAI, Anthropic, Google, Mistral, or Groq
3. Model name — pick a specific model (e.g. gpt-4o-mini)
4. Source locale — your source language (default: en)
5. Target locales — languages to translate into
6. Messages directory — where translation files live (default: ./messages)
7. Include patterns — which source files to scan (auto-detected)
8. i18n library — the import source for useTranslations (keys mode only)
9. Component path — import path for the <T> component (inline mode only)
10. Project context — description for better translations
11. Tone — translation style (professional, casual, etc.)

Generated Files

The wizard creates:

• translate-kit.config.ts — your configuration file

Keys mode

For next-intl projects, it can also create:

• i18n/request.ts — locale detection setup
• Updates to next.config.ts with createNextIntlPlugin

Inline mode

For inline mode, the wizard also:

• Copies t.tsx (client component) and t-server.tsx (server component) to your project
• Creates src/i18n/index.ts with getLocale() and getMessages() helpers
• Wraps your root layout with I18nProvider and wires up locale detection

Full Pipeline

At the end, the wizard offers to run the full pipeline:

scan → codegen → translate

This scans your source code, replaces strings with t() calls, and generates translations in one step.