**Translation infrastructure that sparks joy** ✨
npm install myelinoTranslation infrastructure that sparks joy ✨
Myelin seamlessly integrates localization into your development process, freeing you to prioritize building exceptional user experiences while minimizing time spent on translation tasks.
``bash`
npx myelino --translate
- 🔒 Local First & Private
Translation calls happen on your machine. We never send your data to third-party APIs or external databases.
- 🌍 Open Source
Every line of code is transparent and available on GitHub.
- 🆓 Free Forever
No hidden costs. Now and always.
- 🤖 Multi-Model Support
Choose from openai, deepseek, moonshot, qwen, or bring your own model.
- 📚 First-Class i18n Library Support
Works seamlessly with next-intl, react-intl, react-i18next, and more.
- 🗂 Namespace Support
Supports both namespaced and non-namespaced structures out of the box.
1. Install Myelin
Ensure Node.js (v18+) is installed, then run:
`bash`
npx myelino --init
2. Configure
Create myelin.config.json in your project root:
`json`
{
"aiProvider": "deepseek",
"aiModel": "deepseek-chat",
"locale": {
"location": "locales",
"sourceLanguage": "en",
"targetLanguages": ["es", "fr", "zh", "zh-TW", "en-US"]
}
}
3. Add API Key
Create a .env file with your AI provider key:
`env`
AI_API_KEY=sk-proj-your-api-key
4. Translate!
Run the translation command:
`bash`
npx myelino --translate
Myelin scans your source locale file (e.g., ), translates new/missing strings, and updates target files (e.g., locales/fr.json).
| Key | Description | Example |
| ------------------------ | --------------------------- | ---------------------------- |
| aiProvider | AI service provider | "openai", "deepseek" |aiModel
| | Model ID for the provider | "deepseek-chat", "gpt-4" |locale.location
| | Path to locale directory | "src/locales" |locale.sourceLanguage
| | Base language (ISO 639-1) | "en" |locale.targetLanguages
| | Languages to translate into | ["es", "zh-TW"]` |
- OpenAI
- DeepSeek
- Moonshot
- Qwen
Bring Your Own Model: Custom providers can be integrated via the open-source codebase.
---