mcpField
H
huoshui-pdf-translator

huoshui-pdf-translator

Local
github.com/huoshuiai42Productivity1Updated 59w ago

About huoshui-pdf-translator

huoshui-pdf-translator is an MCP server for translating PDFs, particularly academic papers with mathematical notation. It offers PDF analysis, translation-tool checks, and prompts for paths, options, and troubleshooting.

Key features

  1. 01Translate PDFs while preserving layout
  2. 02Handle mathematical formulas and technical terms
  3. 03Detect languages with Chinese–English specialization
  4. 04Inspect PDF metadata and page count
  5. 05Check translation tool availability
  6. 06Report translation progress

Use cases

  1. 01Translate academic papers with equations
  2. 02Translate technical documents while retaining formatting
  3. 03Inspect a PDF before translation
  4. 04Diagnose translation setup and file errors

MCP support

The MCP server exposes PDF translation, analysis, and status-checking tools, along with PDF information resources and guidance prompts.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"huoshuiai42-huoshui-pdf-translator": {
"command": "uvx",
"args": [
"huoshui-pdf-translator"
]
}
}
}
Run commanduvx huoshui-pdf-translator

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypihuoshui-pdf-translator1.0.0

Tools

4

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
pathstringYesPDF file path (absolute or relative to home directory)

Details

From the project README

Huoshui PDF Translator

Version: 0.1.0
Powered by: FastMCP & PDFMathTranslate-next
PyPI Package: huoshui-pdf-translator

An intelligent PDF translation assistant that specializes in academic papers with mathematical formulas. Built using the FastMCP framework and powered by PDFMathTranslate-next, it provides comprehensive translation capabilities with context-aware assistance.

🌟 Features

Core Translation Capabilities
  • 📚 Academic Papers: Excellent handling of mathematical formulas and equations
  • 🔬 Technical Documents: Preserves formatting and technical terminology
  • 🌐 Multi-language Support: Auto-detection with Chinese ↔ English specialization
  • 🎨 Layout Preservation: Maintains original PDF structure and formatting
Smart Assistant Features
  • 🧠 Context-Aware Prompts: Multiple specialized prompts for different scenarios
  • 🛠️ Tool Status Checking: Verify translation tool installation and availability
  • 📊 PDF Analysis: Get detailed information about PDF files before translation
  • 🔍 Flexible Path Handling: Support for both absolute and relative file paths
  • ⚡ Progress Reporting: Real-time progress updates during translation
  • 🚨 Intelligent Error Handling: Comprehensive error diagnosis and troubleshooting
MCP Features
  • 📋 Resources: Translation capability listings and PDF file information
  • 🎯 Tools: Translation, PDF analysis, and tool status checking
  • 💬 Prompts: Role definitions, path guidance, options explanation, and error troubleshooting
  • 🔒 Security: Safe path validation with system directory protection

🚀 Quick Start

Installation
From MCP Registry (Recommended)

This server is available in the Model Context Protocol Registry. Install it using your MCP client.

mcp-name: io.github.huoshuiai42/huoshui-pdf-translator

Using uvx
uvx huoshui-pdf-translator
Claude Desktop Setup

Add this to your Claude Desktop MCP configuration:

{
  "mcpServers": {
    "huoshui-pdf-translator": {
      "command": "uvx",
      "args": ["huoshui-pdf-translator"]
    }
  }
}
Alternative Installation Methods

Via pipx:

pipx install huoshui-pdf-translator

Via UV tools:

uv tool install huoshui-pdf-translator

Claude Desktop config for UV tools:

{
  "mcpServers": {
    "huoshui-pdf-translator": {
      "command": "uv",
      "args": ["tool", "run", "huoshui-pdf-translator"]
    }
  }
}

📖 Usage

First-Time Setup
  1. Warm up (downloads fonts/models): Use warm_up_translator tool
  2. Check status: Use check_translation_tool tool
  3. Translate: Use translate_pdf tool with your PDF path
MCP Tools
translate_pdf

Translates PDF documents while preserving mathematical formulas and layout.

# Basic usage
translate_pdf(pdf_path="Desktop/paper.pdf")

# With custom output path
translate_pdf(
    pdf_path="Documents/research.pdf",
    output_path="Documents/translated/research_cn.pdf"
)
pdf_get

Retrieves detailed information about a PDF file.

pdf_info = pdf_get(path="Desktop/document.pdf")
# Returns: PDFResource with path, size_bytes, page_count
warm_up_translator

Downloads required assets and models. Run this first to avoid timeouts.

warm_up_translator()
# Downloads fonts and models (~50MB) for faster subsequent translations
check_translation_tool

Verifies PDFMathTranslate-next installation and status.

status = check_translation_tool()
# Returns: status, version, message
MCP Prompts
  • role_and_rules: Core identity and operational rules
  • explain_pdf_paths: Help with file path specifications
  • explain_translation_options: Available options and best practices
  • troubleshoot_translation_error: Error diagnosis and solutions
  • explain_translation_result: Result explanation and next steps
File Path Examples

The assistant supports flexible path specifications:

# Absolute paths
/Users/john/Desktop/research.pdf
C:\Users\John\Documents\paper.pdf

# Relative to home directory
Desktop/research.pdf
Documents/papers/study.pdf

# Simple filenames (assumes home directory)
paper.pdf

🎯 Translation Workflow

  1. Install: uvx huoshui-pdf-translator
  2. Setup Claude Desktop: Add MCP configuration
  3. Warm up: Run warm_up_translator tool (first time only)
  4. Translate: Use translate_pdf with your PDF path
  5. Review: Two files created (dual-language and Chinese-only)

⚡ Performance

  • First translation: 2-5 minutes (downloads fonts/models)
  • Subsequent translations: 30-60 seconds
  • File size limit: 200MB maximum
  • Cache size: ~50MB for fonts and models

🔍 Troubleshooting

Common Issues
Translation Tool Not Available

The tool automatically installs pdf2zh-next when needed. If issues occur:

# Check status
# Use check_translation_tool in Claude Desktop

# Manual install if needed
pip install pdf2zh-next
First Translation Timeout
# Run warmup first
# Use warm_up_translator tool in Claude Desktop
PDF File Not Found
  • Verify file path is correct
  • Use absolute paths for clarity
  • Check file hasn't been moved or deleted
Network Issues
  • Ensure internet connection (required for first-time font downloads)
  • Check firewall settings
Error Diagnosis

The assistant provides intelligent error diagnosis with specific solutions for:

  • File not found errors
  • Invalid PDF files
  • Translation tool issues
  • Network connectivity problems
  • File size limitations

🛠️ Development

For Developers

Install from source:

git clone https://github.com/huoshuiai/huoshui-pdf-translator.git
cd huoshui-pdf-translator
uv sync
uv run python -m huoshui_pdf_translator.main

Build and publish:

uv build
uv run twine upload dist/*
Project Structure
huoshui-pdf-translator/
├── huoshui_pdf_translator/
│   ├── __init__.py      # Package metadata
│   └── main.py         # FastMCP server implementation
├── pyproject.toml      # Package configuration
├── README.md          # This file
└── LICENSE           # Apache-2.0 license

🔄 Updates

Update to latest version:

uvx install --upgrade huoshui-pdf-translator
# or
uv tool upgrade huoshui-pdf-translator

🤝 Contributing

Contributions are welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Add tests if applicable
  5. Submit a pull request

📄 License

This project is licensed under the Apache-2.0 License. See the LICENSE file for details.

🙏 Acknowledgments

  • PDFMathTranslate-next: Core translation engine
  • FastMCP: Framework for intelligent assistant capabilities
  • Anthropic: MCP protocol and ecosystem
  • UV & PyPI: Modern Python packaging and distribution

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.