活水 PDF 转换器 (Huoshui PDF Converter)

A high-quality, cross-platform PDF ↔ Markdown converter implemented as an MCP (Model Context Protocol) server. Supports bidirectional conversion with full Unicode/CJK character support.

Features

Core Capabilities

• PDF → Markdown: Extract text and images with layout preservation
• Markdown → PDF: Generate beautiful PDFs with multiple rendering engines
• Unicode Support: Full support for Chinese, Japanese, Korean, and other Unicode characters
• Cross-Platform: Works on Windows, macOS, and Linux
• MCP Integration: Use with Claude Desktop or any MCP-compatible client

Technical Features

• Pure Python: No external system dependencies required
• Automatic Font Detection: Finds and uses system Unicode fonts
• Smart Engine Selection: Automatically switches engines based on content
• Comprehensive Error Handling: Graceful degradation and detailed logging
• Async Architecture: Non-blocking operations for better performance

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-converter

As a Python Package

pip install huoshui-pdf-converter

Or using uv (recommended):

uv pip install huoshui-pdf-converter

As an MCP Server

Add to your Claude Desktop configuration:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json Linux: ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "huoshui-pdf-converter": {
      "command": "uvx",
      "args": ["huoshui-pdf-converter"],
      "env": {}
    }
  }
}

Or if you prefer to use a specific Python environment:

{
  "mcpServers": {
    "huoshui-pdf-converter": {
      "command": "python",
      "args": ["-m", "huoshui_pdf_converter.server"],
      "env": {}
    }
  }
}

Usage

Command Line Interface

# Convert PDF to Markdown
huoshui-pdf pdf-to-md input.pdf output.md

# Convert Markdown to PDF
huoshui-pdf md-to-pdf input.md output.pdf

# With options
huoshui-pdf md-to-pdf input.md output.pdf --page-size A4 --margin 2cm --font-size 12

As a Python Library

import asyncio
from huoshui_pdf_converter import PDFToMarkdownConverter, MarkdownToPDFConverter

async def main():
    # PDF to Markdown
    pdf_converter = PDFToMarkdownConverter()
    result = await pdf_converter.convert(
        pdf_path="input.pdf",
        output_path="output.md",
        extract_images=True,
        preserve_formatting=True
    )

    # Markdown to PDF
    md_converter = MarkdownToPDFConverter()
    result = await md_converter.convert(
        markdown_path="input.md",
        output_path="output.pdf",
        page_size="A4",
        margin="2cm",
        font_size=12
    )

asyncio.run(main())

MCP Tools

When used as an MCP server, the following tools are available:

1. pdf_to_markdown: Convert PDF files to Markdown

   {
     "pdf_path": "path/to/input.pdf",
     "output_path": "path/to/output.md",
     "extract_images": true,
     "preserve_formatting": true
   }

1. markdown_to_pdf: Convert Markdown files to PDF

   {
     "markdown_path": "path/to/input.md",
     "output_path": "path/to/output.pdf",
     "page_size": "A4",
     "margin": "2cm",
     "font_size": 12
   }

1. list_supported_formats: Get supported formats and engines
2. validate_file: Validate input files before conversion

Supported Formats

Input Formats

• PDF: All standard PDF files (PDF 1.0 - 1.7)
• Markdown: CommonMark and GitHub Flavored Markdown

Output Options

• Page Sizes: A4, A3, Letter, Legal
• Margins: Customizable (e.g., "1cm", "0.5in")
• Font Sizes: Any size in points
• Images: PNG, JPEG extraction from PDFs

Unicode and Font Support

The converter automatically detects and uses appropriate fonts for different languages:

• macOS: Arial Unicode, PingFang SC, STHeiti
• Windows: Microsoft YaHei, SimSun, Arial Unicode MS
• Linux: Noto Sans CJK, Source Han Sans, WenQuanYi

Architecture

Conversion Engines

PDF → Markdown

• PyMuPDF (MuPDF): High-quality text and image extraction

Markdown → PDF

• ReportLab: Best Unicode support, cross-platform compatibility
• xhtml2pdf: Good HTML/CSS rendering (fallback)
• fpdf2: Basic PDF generation (last resort)

Engine Selection Logic

1. Detects CJK characters → Uses ReportLab
2. Complex formatting → Uses xhtml2pdf
3. Basic documents → Uses any available engine

Development

Setup Development Environment

# Clone the repository
git clone https://github.com/yourusername/huoshui-pdf-converter.git
cd huoshui-pdf-converter

# Install dependencies
uv pip install -e ".[dev]"

# Run tests
python test_converter.py

Project Structure

huoshui-pdf-converter/
├── huoshui_pdf_converter/
│   ├── __init__.py
│   ├── server.py           # MCP server implementation
│   ├── pdf_converter.py    # PDF to Markdown converter
│   └── markdown_converter.py # Markdown to PDF converter
├── pyproject.toml
├── README.md
├── LICENSE
└── test_converter.py

Troubleshooting

Common Issues

1. Chinese characters not displaying:

• Ensure Arial Unicode or similar fonts are installed
• The converter will automatically detect and use appropriate fonts

1. Import errors:

• Install all dependencies: pip install huoshui-pdf-converter[all]

1. MCP connection issues:

• Check Claude Desktop logs
• Ensure Python is in your PATH

Logging

Enable debug logging:

import logging
logging.basicConfig(level=logging.DEBUG)

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Built with FastMCP for Model Context Protocol support
• Uses PyMuPDF for PDF parsing
• Uses ReportLab for PDF generation
• Inspired by the need for better PDF ↔ Markdown conversion tools

Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: [email protected]