Configuration Reference

Complete guide to configuring PageTurner for your documentation translation needs.

Basic Configuration

Minimal Setup

from transaurus import DocusaurusMaster

master = DocusaurusMaster(
    # Required parameters
    source_repo_url="https://github.com/your-org/docs",
    target_languages=["es", "fr", "zh-Hans"],

    # Authentication (required)
    github_pat="ghp_xxxxx...",
    anthropic_api_key="sk-ant-xxxxx...",
)

result = master.run_full_translation_pipeline()

Complete Configuration

from transaurus import DocusaurusMaster

master = DocusaurusMaster(
    # Source repository
    source_repo_url="https://github.com/your-org/docs",
    source_language="en",                    # Default: "en"

    # Target languages
    target_languages=["es", "fr", "de", "ja", "zh-Hans"],

    # GitHub configuration
    github_pat="ghp_xxxxx...",
    github_target_org="your-org-i18n",       # Default: "{org}-i18n"

    # AI configuration
    anthropic_api_key="sk-ant-xxxxx...",
    openai_api_key="sk-xxxxx...",            # Optional
    google_api_key="xxxxx...",               # Optional

    # Deployment
    vercel_token="xxxxx...",                 # Optional
)

# Run with options
result = master.run_full_translation_pipeline(
    run_mode="FULL_RUN",                     # or "DRY_RUN", "FAKE_TRANSLATE"
)

Required Parameters

`source_repo_url` (string, required)

GitHub URL of your source Docusaurus repository.

Format: https://github.com/{org}/{repo}

Example:

source_repo_url="https://github.com/facebook/docusaurus"

`target_languages` (list, required)

List of target language codes (ISO 639-1).

Format: List of 2-letter language codes

Example:

target_languages=["es", "fr", "de", "ja"]

`github_pat` (string, required)

GitHub Personal Access Token with repo scope.

How to get:

Go to GitHub Settings → Tokens
Check repo scope
Generate token

Example:

github_pat="ghp_1A2B3C4D5E6F7G8H9I0J"

`anthropic_api_key` (string, required)

Anthropic API key for Claude AI.

How to get:

Visit Anthropic Console
Create API key
Add credits ($5+ recommended)

Example:

anthropic_api_key="sk-ant-api03-xxx..."

Optional Parameters

`source_language` (string, optional)

Source language code. Default: "en"

Example:

source_language="en"  # English

`github_target_org` (string, optional)

GitHub organization for translated repositories.

Default: {source_org}-i18n

Example:

# Source: https://github.com/facebook/docusaurus
# Default target org: "facebook-i18n"
# Custom:
github_target_org="facebook-translations"

`vercel_token` (string, optional)

Vercel API token for automatic deployment.

How to get:

Go to Vercel Settings → Tokens
Create token

Example:

vercel_token="1234567890abcdef"

Note: Without Vercel token, translations are created on GitHub but not automatically deployed.

`openai_api_key` (string, optional)

OpenAI API key for GPT models (used for specialized tasks).

Example:

openai_api_key="sk-proj-xxx..."

`google_api_key` (string, optional)

Google API key for Gemini models (used for multimodal content).

Example:

google_api_key="AIzaSy..."

Supported Languages

PageTurner supports 100+ languages via Claude AI. Common language codes:

Most Popular

Language	Code	Example Translation Time (100 pages)
Spanish	`es`	20-30 min
French	`fr`	20-30 min
German	`de`	20-30 min
Japanese	`ja`	25-35 min
Chinese (Simplified)	`zh-Hans`	25-35 min
Chinese (Traditional)	`zh-Hant`	25-35 min
Korean	`ko`	25-35 min
Portuguese	`pt`	20-30 min
Russian	`ru`	25-35 min
Arabic	`ar`	30-40 min

European Languages

["es", "fr", "de", "it", "pt", "ru", "pl", "nl", "sv", "da", "no", "fi"]

Asian Languages

["ja", "ko", "zh-Hans", "zh-Hant", "th", "vi", "id", "ms"]

All Supported Languages

See ISO 639-1 language codes for complete list.

Note: Translation quality varies by language. Major languages (es, fr, de, ja, zh) average 91+/100. Less common languages may be 85-90/100.

Run Modes

PageTurner supports three execution modes:

1. `FULL_RUN` (Default)

Complete translation and deployment pipeline.

What it does:

✅ Clones source repository
✅ Analyzes and segments content
✅ Translates using AI (costs API credits)
✅ Creates mirror repositories on GitHub
✅ Deploys to Vercel (if token provided)

Usage:

result = master.run_full_translation_pipeline(run_mode="FULL_RUN")

When to use: Production translation with real deployment

2. `DRY_RUN`

Planning and validation mode - no LLM calls or deployments.

What it does:

✅ Clones source repository
✅ Analyzes and segments content
✅ Shows what would be translated
❌ No AI translation (no API costs)
❌ No repository creation
❌ No deployment

Usage:

result = master.run_full_translation_pipeline(run_mode="DRY_RUN")

When to use:

Estimate translation costs before committing
Validate source repository is compatible
Check how many segments will be translated
Test configuration without spending credits

Output:

📊 DRY RUN Summary:
- Source repository: validated ✓
- Translatable segments: 156
- Estimated tokens: 45,000
- Estimated cost: $4.50
- Target languages: es, fr, zh-Hans
- Repository structure: valid ✓

3. `FAKE_TRANSLATE`

Full pipeline testing with fake translations (uses source language as "translation").

What it does:

✅ Clones source repository
✅ Analyzes and segments content
✅ "Translates" (copies source as translation)
✅ Creates mirror repositories on GitHub
✅ Deploys to Vercel (if token provided)
⚠️ Minimal API costs (no real translation)

Usage:

result = master.run_full_translation_pipeline(run_mode="FAKE_TRANSLATE")

When to use:

Test deployment pipeline without translation costs
Verify repository creation works
Test Vercel integration
Validate CI/CD workflows
Debug issues without spending credits

Result: Deployed sites in "Spanish", "French", etc. but all content is still in English.

Repository Organization

Default Structure

Source repository:
  https://github.com/your-org/docs

Mirror repositories (default):
  https://github.com/your-org-i18n/docs-es
  https://github.com/your-org-i18n/docs-fr
  https://github.com/your-org-i18n/docs-zh-hans

Custom Organization

master = DocusaurusMaster(
    source_repo_url="https://github.com/your-org/docs",
    github_target_org="your-org-translations",
    target_languages=["es", "fr"],
)

# Creates:
# https://github.com/your-org-translations/docs-es
# https://github.com/your-org-translations/docs-fr

Repository Naming Convention

Format: {repo-name}-{language-code}

Examples:

docs → docs-es, docs-fr, docs-ja
documentation → documentation-es, documentation-fr
my-awesome-docs → my-awesome-docs-es, my-awesome-docs-fr

Deployment Options

Automatic Vercel Deployment

With Vercel token, PageTurner automatically:

Creates Vercel project
Links to GitHub repository
Configures build settings
Deploys to production
Returns deployment URL

Configuration:

master = DocusaurusMaster(
    # ... other config
    vercel_token="your_vercel_token",
)

result = master.run_full_translation_pipeline()

print(result['vercel_deployment_urls'])
# {
#   'es': 'https://docs-es.vercel.app',
#   'fr': 'https://docs-fr.vercel.app',
# }

Manual Deployment

Without Vercel token, you can deploy manually:

Option 1: Import to Vercel

Go to Vercel
Click "New Project"
Import from GitHub
Select your-org-i18n/docs-es
Deploy

Option 2: Netlify

Go to Netlify
Add new site
Connect to GitHub
Select repository
Deploy

Option 3: GitHub Pages

Enable GitHub Pages in repository settings
Configure build workflow
Deploy to gh-pages branch

Advanced Configuration

Custom Configuration File (Optional)

Create transaurus.config.json in your project:

{
  "source_repo_url": "https://github.com/your-org/docs",
  "target_languages": ["es", "fr", "de", "ja", "zh-Hans"],
  "source_language": "en",
  "github_target_org": "your-org-i18n",
  "deployment": {
    "platform": "vercel",
    "auto_deploy": true
  },
  "translation": {
    "quality_threshold": 85,
    "retry_failed_segments": true
  }
}

Load configuration:

import json
from transaurus import DocusaurusMaster

with open("transaurus.config.json") as f:
    config = json.load(f)

master = DocusaurusMaster(**config)

Cost Estimation

Translation Costs

Costs depend on:

Number of pages
Content length
Number of target languages
Translation memory reuse

Typical costs (using Claude Sonnet):

Pages	Languages	First Translation	Updates (with TM)
50	3	$15	$2-3
100	3	$30	$3-5
200	3	$60	$6-10
100	10	$100	$10-15

Formula: ~$0.10 per page per language (first translation)

Cost Optimization

1. Use Translation Memory

First translation: $30 for 100 pages × 3 languages
Update 10 pages: $3 (90% savings)

2. Use DRY_RUN First

# Estimate costs before running
result = master.run_full_translation_pipeline(run_mode="DRY_RUN")
print(f"Estimated cost: ${result['estimated_cost']}")

3. Translate Incrementally

Translate high-priority pages first
Add languages gradually
Use translation memory across projects

Troubleshooting

Configuration Validation

PageTurner validates configuration on startup:

✓ Source repository accessible
✓ GitHub PAT has required permissions
✓ Anthropic API key valid
✓ Target languages supported
✓ Repository structure valid

Common Issues

"Invalid source repository"

Ensure repository is a valid Docusaurus site
Check docusaurus.config.js exists
Verify repository is accessible with GitHub PAT

"Insufficient GitHub permissions"

GitHub PAT must have repo scope
Regenerate token if needed

"Anthropic API error"

Check API key is valid
Ensure sufficient credits ($5+ recommended)
Verify rate limits not exceeded

Next Steps

Run your first translation - Quick getting started guide
Understand the pipeline - How PageTurner works
Check FAQ - Common questions and troubleshooting

Questions? Contact us or book a demo.

Basic Configuration​

Minimal Setup​

Complete Configuration​

Required Parameters​

source_repo_url (string, required)​

target_languages (list, required)​

github_pat (string, required)​

anthropic_api_key (string, required)​

Optional Parameters​

source_language (string, optional)​

github_target_org (string, optional)​

vercel_token (string, optional)​

openai_api_key (string, optional)​

google_api_key (string, optional)​

Supported Languages​

Most Popular​

European Languages​

Asian Languages​

All Supported Languages​

Run Modes​

1. FULL_RUN (Default)​

2. DRY_RUN​

3. FAKE_TRANSLATE​

Repository Organization​

Default Structure​

Custom Organization​

Repository Naming Convention​

Deployment Options​

Automatic Vercel Deployment​

Manual Deployment​

Advanced Configuration​

Custom Configuration File (Optional)​

Cost Estimation​

Translation Costs​

Cost Optimization​

Troubleshooting​

Configuration Validation​

Common Issues​

Next Steps​

Basic Configuration

Minimal Setup

Complete Configuration

Required Parameters

`source_repo_url` (string, required)

`target_languages` (list, required)

`github_pat` (string, required)

`anthropic_api_key` (string, required)

Optional Parameters

`source_language` (string, optional)

`github_target_org` (string, optional)

`vercel_token` (string, optional)

`openai_api_key` (string, optional)

`google_api_key` (string, optional)

Supported Languages

Most Popular

European Languages

Asian Languages

All Supported Languages

Run Modes

1. `FULL_RUN` (Default)

2. `DRY_RUN`

3. `FAKE_TRANSLATE`

Repository Organization

Default Structure

Custom Organization

Repository Naming Convention

Deployment Options

Automatic Vercel Deployment

Manual Deployment

Advanced Configuration

Custom Configuration File (Optional)

Cost Estimation

Translation Costs

Cost Optimization

Troubleshooting

Configuration Validation

Common Issues

Next Steps