
i18n-agent
ai.i18nagentPublished by i18nagent.aiOfficial publisherv1.16.3Updated Oct 24, 2026
AI-powered translation for 48 languages with context-aware quality
Overview
Lets an assistant translate text and localization files into 48 languages through a paid, credit-based translation service.
- What you can do
- Exposes tools to translate text and files (JSON, YAML, CSV, XML, Markdown, properties, PDF, Word, Gettext) while preserving structure, with context options for audience, industry and region. It also analyzes content for translation readiness, lists supported languages with quality ratings, checks credit balance and word estimates, and tracks asynchronous jobs for large files.
- When to use it
- Useful when localizing app strings, documentation or content files and you want the assistant to translate them in place with context-aware quality. Also helpful for pre-checking content quality and estimating cost before spending credits.
- Requirements
- A remote endpoint at mcp.i18nagent.ai or the local npm package @i18n-agent/mcp-client run with Node.js. An API key from the provider's dashboard, supplied as the I18N_AGENT_API_KEY environment variable, and network access. The MCP_SERVER_URL variable can point at the server URL.
Installation
In SourceWeft
- Open i18n-agent in the dashboard and add it to a workspace.
- Connect the required credentials. They are stored privately and never shown on public pages.
- Enable the server for the chats that should use its tools.
Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.
Use with an agent
Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.
Read https://sourceweft.com/mcp/ai.i18nagent%2Fi18n-agent/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.Other MCP clients
Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.
{
"mcpServers": {
"i18n-agent": {
"type": "http",
"url": "https://mcp.i18nagent.ai/mcp"
}
}
}README
🌍 i18n-agent MCP Client
Professional translation service client for Claude, Cursor, VS Code, Antigravity, and other AI IDEs using the Model Context Protocol (MCP).
✨ Features
- 🎯 Smart Translation: Context-aware translations with cultural adaptation
- 📁 File Translation: Support for JSON, YAML, CSV, XML, Markdown, and more
- ⚡ Large File Support: Async processing for files >50KB with progress tracking
- 🔄 Timeout Improvements: Extended timeouts (5-10 min) for large translations
- 📊 Progress Tracking: Real-time job status and completion monitoring
- 💰 Credit Tracking: Real-time credit balance and word count estimates
- 🌐 48 Languages: Comprehensive language support with regional variants
- 🔧 Easy Setup: One-command installation for major AI IDEs
🚀 Quick Installation
The installer will detect all available AI IDEs and configure them automatically.
Claude Code Marketplace Installation
For Claude Code users, you can install directly from the marketplace:
Step 1: Get your API key
- Visit app.i18nagent.ai
- Sign up or log in
- Copy your API key (starts with "i18n_")
Step 2: Set environment variable
Replace your-api-key-here with your actual API key.
Step 3: Install from marketplace
Step 4: Restart Claude Code
That's it! The plugin will automatically use your API key from the environment variable.
🔑 Setup API Key
-
Get your API key from app.i18nagent.ai
-
Set environment variable:
-
Make it permanent (add to ~/.bashrc or ~/.zshrc):
-
Restart your AI IDE to load the new configuration
🎮 Usage Examples
Text Translation
File Translation
Credit Check
Language Support
Content Analysis
🛠 Supported AI IDEs
Note: The installer automatically detects your platform and uses the correct config paths.
🌐 Language Support (48 Languages)
- bg: Bulgarian
- ca: Catalan
- cs: Czech
- da: Danish
- de: German
- el: Greek
- en: English
- en-AU: English (Australia)
- en-CA: English (Canada)
- en-GB: English (United Kingdom)
- en-US: English (United States)
- es: Spanish
- es-MX: Spanish (Mexico)
- et: Estonian
- fi: Finnish
- fr: French
- fr-CA: French (Canada)
- hi: Hindi
- hr: Croatian
- hu: Hungarian
- id: Indonesian
- is: Icelandic
- it: Italian
- ja: Japanese
- ko: Korean
- lt: Lithuanian
- lv: Latvian
- ms: Malay
- nl: Dutch
- no: Norwegian
- pl: Polish
- pt: Portuguese
- pt-BR: Portuguese (Brazil)
- ro: Romanian
- ru: Russian
- sk: Slovak
- sl: Slovenian
- sr: Serbian
- sv: Swedish
- th: Thai
- tl: Filipino
- tr: Turkish
- uk: Ukrainian
- vi: Vietnamese
- zh-Hans: Chinese (Simplified)
- zh-Hant-HK: Chinese (Traditional, Hong Kong)
- zh-Hant-TW: Chinese (Traditional, Taiwan)
📁 Supported File Formats
🔧 Manual Setup
If auto-installation fails, you can manually configure your IDE:
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
Cursor / VS Code
Create .cursor/mcp_settings.json or .vscode/mcp_settings.json:
💡 Usage Tips
Translation Context
- Target Audience: Specify "technical", "casual", "formal", or "general"
- Industry Context: Use "technology", "healthcare", "finance", "education"
- Regional Variations: Add regions like "Spain", "Mexico", "Brazil"
File Translation
- Preserve Structure: Keeps original file format and structure
- Output Format: Convert between formats (JSON ↔ YAML ↔ CSV)
- Large Files: Automatically chunks large files for processing
- Async Processing: Files >50KB processed asynchronously with job tracking
- Progress Monitoring: Real-time status updates for long-running translations
- Timeout Resilience: Up to 10 minutes for large translation jobs
Large Translation Handling
- Async Processing: >100 texts or >50KB files processed asynchronously
- Job Tracking: Unique job IDs for monitoring long-running translations
- Progress Updates: Real-time completion percentages and status
- Extended Timeouts: 5-10 minute timeouts prevent interruptions
- Automatic Polling: Client automatically polls for job completion
Credit Management
- Cost: 1 credit per word ($0.01/word)
- Monitoring: Check balance before large translations
- Estimates: Get word count estimates before translation
Quality Warnings
- Source Analysis: By default, source content is analyzed for quality issues before translation
- Skip Warnings: Use
skipWarnings: trueto bypass warnings in automated workflows - Trade-off: Skipping warnings may reduce translation quality as potential issues aren't addressed
- Best Practice: Keep warnings enabled (default) for production translations
🚨 Troubleshooting
Installation Issues
Permission denied:
IDE not detected:
MCP Connection Issues
"Failed" status in Claude Code:
This usually happens with Node Version Managers (nvm, fnm, n). The installer now automatically detects nvm and creates a wrapper script. If you still have issues:
-
Check your Node installation:
-
Manual wrapper script (if auto-detection fails): Create
~/.claude/run-mcp.sh:Make it executable:
-
Update Claude configuration: Edit
~/.claude.json: -
Restart Claude Code completely (not just close window, quit the app)
Runtime Issues
API Key not found:
Connection errors:
- Check your internet connection
- Verify API key is valid
- Try again after a few seconds
Translation quality:
- Use Tier 1 languages for production
- Add context with industry/audience parameters
- Review Tier 2/3 translations manually
📊 Pricing
- Pay-per-use: 1 credit per word ($0.01/word)
- No subscriptions: Only pay for what you translate
- Bulk discounts: Available for enterprise usage
- Free tier: New accounts get starter credits
🔐 Privacy & Security
- No data storage: Translations are processed in real-time
- Encrypted transport: All data sent over HTTPS
- API key security: Keys are stored locally, never transmitted in logs
- GDPR compliant: EU privacy standards
🏗️ Architecture
The MCP client installer uses a modular architecture for bulletproof installation:
Key Features
- Atomic Config Writes: Use temp files + rename for crash safety
- Automatic Backups: Create timestamped backups before modifications
- Transport Detection: Auto-detect SSE vs stdio from config
- Comprehensive Validation: Validate all inputs at module boundaries
- Zero External Dependencies: Uses only Node.js built-ins
Architecture Details
See detailed documentation:
- Module Architecture - Module responsibilities and design
- Full Architecture Guide - Complete system design
🤝 Contributing
We welcome contributions! Please see our Contributing Guidelines.
Development Setup
Publishing
Use the publishing script with built-in test gate:
📝 License
MIT License - see LICENSE file for details.
Copyright (c) 2025 FatCouple OÜ
🔗 Links
- Website: i18nagent.ai
- Dashboard: app.i18nagent.ai
- Documentation: docs.i18nagent.ai
- GitHub: github.com/i18n-agent/mcp-client
- Issues: github.com/i18n-agent/mcp-client/issues
🆘 Support
- Email: [email protected]
- Documentation: docs.i18nagent.ai
🔧 Available MCP Tools
translate_text
Translate text content with cultural adaptation and context awareness.
Parameters:
texts(array): Array of strings to translatetargetLanguage(string): Target language codetargetAudience(string): Target audience contextindustry(string): Industry contextnamespace(string, optional): Optional namespace identifier for backend tracking and project organizationsourceLanguage(string, optional): Source language (auto-detected if not provided)region(string, optional): Specific region for localizationskipWarnings(boolean, optional): Skip source text quality warnings (default: false). ⚠️ WARNING: May hurt translation quality by bypassing source analysis. Only use when confident about content quality or in automated workflows.
translate_file
Translate files while preserving structure and format.
Parameters:
filePathorfileContent(string): File path or content to translatefileType(string): File format (json, yaml, xml, csv, txt, md, etc.)targetLanguage(string): Target language codenamespace(string, required): Unique namespace identifier for backend tracking and project organizationpreserveKeys(boolean): Whether to preserve object keys/structureoutputFormat(string): Output format (same, json, yaml, txt)skipWarnings(boolean, optional): Skip source text quality warnings (default: false). ⚠️ WARNING: May hurt translation quality by bypassing source analysis. Only use when confident about content quality or in automated workflows.
analyze_content
Analyze content for translation readiness and get improvement suggestions before translation. This helps identify potential issues and optimize content before spending credits on translation.
Parameters:
content(string/array/object): Content to analyzetargetLanguage(string): Target language for translationfileType(string, optional): File type if content is from a filesourceLanguage(string, optional): Source language (auto-detected)industry(string): Industry contexttargetAudience(string): Target audienceregion(string, optional): Specific region for localization
Returns:
- Source language detection with confidence score
- Content type and tone analysis
- Translation readiness score (0-100)
- Specific improvement suggestions
- Quality metrics and issues
- Warnings for potential problems
- Estimated credits required
list_supported_languages
Get list of all supported languages with quality ratings.
Parameters:
includeQuality(boolean): Include quality ratings (default: true)
get_credits
Check remaining translation credits and word count estimates.
Parameters:
apiKey(string): Your API key
check_translation_status
Check status of async translation jobs (for large files).
Parameters:
jobId(string): Job ID from async translation
Made with ❤️ by FatCouple OÜ
Source: README.md at commit 320290a
Tools
0Version history
1- v1.16.3LatestSep 16, 2026
Score
85% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 daysRequired25 pts
- Auth declaredRequired10 pts
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts