Changelog

All notable changes to i18nizer are documented here.

0.7.1 - Pluralization and Rich Text Support

Release Date: January 2026

✨ Added

  • Automatic Pluralization Detection: i18nizer now automatically detects and converts pluralization patterns

    • Detects ternary patterns like count === 1 ? 'item' : 'items'
    • Automatically generates ICU message format translations with plural rules
    • Replaces patterns with t("key", { count }) calls
    • Preserves variable names in placeholders (e.g., totalItems, userCount)
    • Works in both JSX children and attributes
    • Supports ===, ==, and both count === 1 and 1 === count patterns

  • Automatic Rich Text Formatting: i18nizer now automatically handles JSX elements within text

    • Detects mixed text and inline elements (e.g., <a>, <strong>, <em>, <span>)
    • Automatically generates t.rich() calls with element mappings
    • Preserves element structure: a: (chunks) => <a>{chunks}</a>
    • Handles multiple inline elements in the same text
    • Generates proper ICU MessageFormat with XML-style tags like <a>link text</a>
    • Compatible with both next-intl's t.rich() and i18next's Trans component

  • Comprehensive Documentation: Added guides and examples for automatic transformations

    • Before/after code samples showing automatic detection and conversion
    • Integration examples with i18next and next-intl
    • Real-world usage scenarios and patterns
    • Support for multiple plural categories (zero, one, two, few, many, other)

🔄 Changed

  • Enhanced README with automatic pluralization and rich text transformation examples
  • Updated documentation site with detailed automatic conversion guides
  • Improved examples page with real-world automatic transformation scenarios

📝 Notes

  • Pluralization and rich text are automatically detected and converted during extraction
  • Generated translations use ICU MessageFormat compatible with i18next and next-intl
  • No manual configuration or refactoring required—works out of the box
  • Both features are fully tested with comprehensive test suites

Documentation:

0.7.0 - AI Provider and Paths Configuration

Release Date: January 2026

✨ Added

  • AI Provider Configuration: Configure AI provider and model directly in i18nizer.config.yml

    • ai.provider: Choose between openai, gemini, or huggingface (default: openai)
    • ai.model: Specify the model to use (default: gpt-4)
    • Provider validation ensures only supported providers are accepted
    • CLI --provider flag still works and overrides config setting

  • Paths Configuration: Define default project paths in config

    • paths.src: Source directory (default: src)
    • paths.i18n: i18n output directory (default: i18n)

🔄 Changed

  • AI provider now defaults to openai instead of huggingface when not specified
  • generateTranslations() function now accepts optional model parameter for custom model selection

🐛 Fixed

  • Config merging now properly handles optional ai and paths fields

📝 Notes

  • Fully backward compatible - existing configs without ai or paths fields will use default values
  • Configuration example:
Terminal
ai:
  provider: gemini
  model: gemini-2.5-flash

paths:
  src: source
  i18n: locales

Documentation:

0.6.2 - Minor Fixes

  • Bug fixes and stability improvements

0.6.1 - Some Improvements

✨ Added

  • messages.generated.ts aggregator file consolidating all JSON translations by language
  • regenerate CLI command to rebuild messages.generated.ts automatically
  • Automatic addition of i18nizer.config.yml to .gitignore if present
  • Option to use aggregator as single import point instead of multiple dynamic imports

🔄 Changed

  • JSON file naming convention standardized to lowercase with hyphens (e.g., notification-item.json) for valid TypeScript imports
  • Auto-generated aggregator now sanitizes identifiers for TypeScript, converting JSON filenames into valid import identifiers
  • Namespaces inside JSON remain in PascalCase, keys remain camelCase

🐛 Fixed

  • Resolved invalid import errors caused by hyphens in JSON filenames
  • Eliminated the need for multiple dynamic imports per component or page

📝 Notes

All messages for a given locale can now be accessed via messages.generated.ts:

Terminal
import { messages } from './messages.generated';

const localeMessages = messages[locale];

This consolidates translations, simplifies imports, and improves performance.

0.6.0 - Opinionated System & Generated Aggregator

This release introduces an opinionated system with automatic aggregator generation and enhanced configuration options.

✨ New Features

  • Auto-Generated Aggregator (/i18n/messages.generated.ts)

    • Automatically scans /messages/<locale>/ directories and generates a TypeScript aggregator file
    • Imports all translation JSON files and exports them as a unified messages object
    • Compatible with i18next and next-intl
    • Includes warning comment to prevent manual editing
    • Regenerates automatically after translate and extract commands
    • Displays statistics about processed locales and namespaces

  • Automatic .gitignore Management

    • .i18nizer/ folder is automatically added to .gitignore during i18nizer start
    • Keeps cache and temporary files out of version control
    • i18nizer.config.yml remains tracked in git for team collaboration
    • Prevents duplicate entries with intelligent pattern detection
    • Handles various gitignore patterns (exact match, leading slash, wildcard)

  • Configurable Locales

    • New locales field in i18nizer.config.yml under messages section
    • Define project locales once in config instead of passing --locales flag every time
    • Commands still accept --locales flag to override config setting
    • Default locales: ["en", "es"]

🏗️ Opinionated Structure

  • Messages Directory: /messages/<lang>/<namespace>.json

    • Separate folders per language for easier versioning
    • Namespace = PascalCase (component name)
    • Keys = camelCase
    • No language mixing in files

  • Generated Aggregator: /i18n/messages.generated.ts

    • Spread operator pattern for compatibility
    • Sorted imports for consistency
    • Auto-regenerated, never edited manually

🧪 Testing

  • Added 15 new comprehensive tests
    • 7 tests for aggregator generation (edge cases, sorting, multiple locales)
    • 8 tests for .gitignore management (patterns, duplicates, empty files)
  • All 118 tests passing with no regressions

0.5.1 - Batch Key Generation

  • Implement batch key generation to minimize AI requests
  • Minor fixes

0.5.0 - AI Key Generation Improvements

✨ New Features

  • AI-Powered Key Generation
  • Intelligent Caching
  • Graceful Fallback

0.4.0 - Project Integration

This release completes Phase 1 - project-level integration with configuration, caching, and intelligent string management.

✨ New Features

  • i18nizer start command: Initialize i18nizer in your project

    • Interactive mode with framework and i18n library detection
    • Automatic framework detection (Next.js, React)
    • Automatic i18n library detection (next-intl, react-i18next, i18next)
    • Manual configuration with --framework and --i18n flags
    • Non-interactive mode with --yes flag for CI/automation
    • Framework presets with optimal defaults
    • Generates i18nizer.config.yml configuration file
    • Creates .i18nizer/ directory for caching
    • Sets up configurable messages/ directory

  • i18nizer translate command: Enhanced translation workflow

    • Project-level mode with --all flag to translate all components
    • Single-file mode for targeted translation
    • --dry-run flag to preview changes
    • --show-json flag to display generated translation output
    • Integrated with project configuration system

  • Configuration system:

    • YAML-based configuration (i18nizer.config.yml)
    • Framework presets (Next.js, React, Custom)
    • I18n library-specific configurations
    • Combined presets (e.g., Next.js + next-intl)
    • autoInjectT option to control automatic translation function injection
    • Configurable i18n function and import source
    • Customizable messages directory and default locale
    • Configurable behavior (deduplication, allowed props/functions)

  • Intelligent caching:

    • Translation cache in .i18nizer/cache/
    • Avoids redundant AI translation requests
    • String-to-key mapping for fast lookups
    • Persistent across multiple runs

  • String deduplication:

    • Detects duplicate strings across components
    • Reuses deterministic keys for identical strings
    • Reduces translation costs and ensures consistency

  • Enhanced CLI feedback:

    • Colored output with chalk
    • Progress spinners with ora
    • Interactive prompts with inquirer
    • Detailed summary statistics
    • Clear error messages and guidance

0.3.0 - Foundation & Reliability

This release completes Phase 0 - stabilizing the core extraction, replacement, and reliability features.

✨ New Features

  • Expanded prop support: aria-placeholder, label, text, tooltip, helperText
  • Deterministic key generation with stable, human-readable camelCase keys
  • Improved JSON output quality:
    • Alphabetically sorted keys for stable output
    • Consistent 2-space indentation
    • Trailing newlines added to all JSON files
  • Enhanced non-translatable filtering

🧪 Testing

  • Expanded test suite from 26 to 60+ tests
  • Comprehensive tests for:
    • All new supported props
    • Non-translatable content filtering
    • Deterministic key generation
    • JSON output quality and stability
  • All tests passing on Node 18, 20, and latest

0.2.0 - Enhanced Text Extraction

  • Enhanced text extraction for complex JSX expressions:

    • Support for ternary operators in JSX attributes and children
    • Support for logical AND (&&) and OR (||) operators
    • Support for curly-braced string literals in JSX attributes
    • Improved filtering to skip non-translatable content

  • Improved test coverage:

    • Comprehensive test suite for all extraction edge cases
    • Tests verify ternary operators, logical operators, nested expressions

  • CI/CD improvements:

    • Test status badge added to README
    • Test workflow runs automatically on every push

0.1.5 - Minor Fixes

  • Bug fixes and stability improvements

0.1.4 - Member Function Support

  • Support for member function calls (e.g., toast.error, toast.success)
  • getFullCallName helper added for member expressions
  • Template literals inside member functions now processed correctly
  • Placeholders in t("key", { ... }) now supported
  • Comments updated to English for clarity

0.1.3 - TSX/JSX Parsing Fixes

  • Fixed TSX/JSX parsing to work with any file path
  • Avoided Windows EPERM errors when scanning system folders
  • Updated parseFile to use minimal in-memory ts-morph project
  • Fixed extraction of translation keys
  • Updated buildPrompt for better AI outputs

0.1.2 - Minor Fixes

  • Bug fixes and stability improvements

0.1.1 - Project Structure

  • Added .i18nizer folder in project root
    • .i18nizer/tsconfig.json for isolated TypeScript analysis
    • .i18nizer/api-keys.json for storing API keys locally
    • .i18nizer/messages/ for generated locale JSON files
  • Automatically creates default configs if they don't exist
  • Ensures CLI works independently of user project tsconfig

0.1.0 - Initial Release

  • Initial release of i18nizer
  • Supported AI providers: OpenAI, Hugging Face, Gemini
  • CLI works with JSX and TSX files
  • Console logs show paths of saved locale files
  • Works independently of user project tsconfig

For the latest updates, visit the GitHub repository.