Start

Translations & Localization

How multilingual outputs are generated and organized.

Overview

The system supports the target languages configured in data/config/languages.yaml, with script-aware quality checks and evidence-status metadata.

Run Translation

From the repository root:

uv run start-curriculum --non-interactive --stages translations \
  --languages Spanish French Chinese --json

If --languages is omitted, languages are read from data/config/languages.yaml.

Outputs

Artifacts are organized under data/translated_curriculums/ by language:

data/translated_curriculums/
├── spanish/
├── french/
├── chinese/
└── tagalog/

Each language folder mirrors the source curriculum file structure and filenames.

Configuration

Languages are defined in:

data/config/languages.yaml

Add or remove language keys as needed. The translation script validates requested languages against this file.

Tips

Prompt Template

Translation guidance lives in:

data/prompts/translation.md

This template ensures tone, terminology, and cultural appropriateness while preserving technical accuracy.