voidlight-markdown-mcp-server

VoidLight00/voidlight-markdown-mcp-server

3.2

If you are the rightful owner of voidlight-markdown-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

Voidlight Markdown MCP Server is an enhanced version of Microsoft MarkItDown, designed to support Korean language and additional file formats.

Tools
  1. convert_to_markdown

    Converts various file formats to markdown.

  2. analyze_document_structure

    Analyzes document structure and extracts metadata.

  3. list_supported_formats

    Lists all supported file formats.

VoidLight MarkItDown MCP Server

PyPI version Docker License: MIT Tests

Enhanced MarkItDown MCP Server with Korean support and advanced document conversion capabilities for Claude Desktop.

🚀 Quick Install

pip install voidlight-markitdown-mcp

⚡ Setup for Claude Desktop

1. Install and Setup

# Install the package
pip install voidlight-markitdown-mcp

# Auto-configure Claude Desktop
voidlight-markitdown-mcp --setup

2. Manual Configuration

Add to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "voidlight-markitdown": {
      "command": "voidlight-markitdown-mcp"
    }
  }
}

3. Start Using

  1. Restart Claude Desktop
  2. Use the convert_to_markdown tool to convert documents
  3. Support for 50+ file formats including:
    • 📄 PDF, DOCX, PPTX, TXT
    • 🌐 HTML, RTF, CSV, Excel
    • 🖼️ Images with OCR (Korean supported)
    • 🎵 Audio transcription
    • 📱 Mobile formats (HWP, etc.)

📦 Installation Options

Core Installation (Recommended)

pip install voidlight-markitdown-mcp

With All Features

pip install "voidlight-markitdown-mcp[all]"

Specific Feature Sets

# Office documents (Word, Excel, PowerPoint)
pip install "voidlight-markitdown-mcp[office]"

# PDF processing
pip install "voidlight-markitdown-mcp[pdf]"

# Image OCR
pip install "voidlight-markitdown-mcp[image]"

# Audio transcription  
pip install "voidlight-markitdown-mcp[audio]"

# Korean language support
pip install "voidlight-markitdown-mcp[korean]"

# Cloud services
pip install "voidlight-markitdown-mcp[cloud]"

🐳 Docker Usage

Quick Start

docker run -v $(pwd)/documents:/workdir ghcr.io/voidlight00/voidlight-markdown-mcp-server

Docker Compose

# Clone repository
git clone https://github.com/VoidLight00/voidlight-markdown-mcp-server.git
cd voidlight-markdown-mcp-server

# Start services
docker-compose up -d

# Process documents
docker-compose exec voidlight-markitdown-mcp voidlight-markitdown-mcp --status

🛠️ CLI Usage

Check Installation

voidlight-markitdown-mcp --status

Install Additional Dependencies

voidlight-markitdown-mcp --install-deps all

Setup Claude Desktop

voidlight-markitdown-mcp --setup

🌟 Features

Enhanced Document Support

  • 50+ File Formats: PDF, DOCX, PPTX, Excel, HTML, RTF, CSV, TXT, Images
  • Korean Optimization: Enhanced Korean text processing and OCR
  • Cloud Integration: Azure Document Intelligence, Google Cloud Vision
  • Audio Processing: Speech-to-text with Korean support
  • Image OCR: Text extraction from images with multilingual support

Performance & Reliability

  • Memory Efficient: Streaming processing for large files
  • Error Handling: Graceful fallbacks and detailed error messages
  • Caching: Smart caching for repeated conversions
  • Parallel Processing: Multi-threaded conversion for batch operations

Developer Experience

  • Simple API: One-command installation and setup
  • Comprehensive Logging: Detailed conversion logs and metrics
  • Extensible: Plugin architecture for custom converters
  • Type Safety: Full TypeScript/Python type annotations

📋 Supported Formats

CategoryFormats
DocumentsPDF, DOCX, PPTX, RTF, TXT, MD
SpreadsheetsXLSX, XLS, CSV
WebHTML, XML, RSS
ImagesPNG, JPG, GIF, BMP, TIFF
AudioMP3, WAV, M4A, OGG
KoreanHWP, OCR with Korean
ArchivesZIP (extract and convert)

🔧 Configuration

Environment Variables

export MCP_WORKDIR=/path/to/workspace
export MCP_LOG_LEVEL=INFO
export AZURE_DOCUMENT_INTELLIGENCE_ENDPOINT=your_endpoint
export OPENAI_API_KEY=your_api_key

Configuration File

Create ~/.voidlight-markitdown-mcp/config.json:

{
  "conversion": {
    "include_metadata": true,
    "extract_images": false,
    "korean_optimization": true
  },
  "performance": {
    "max_workers": 4,
    "memory_limit": "1GB"
  }
}

🚀 Development

Setup Development Environment

git clone https://github.com/VoidLight00/voidlight-markdown-mcp-server.git
cd voidlight-markdown-mcp-server

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

# Run tests
pytest

# Run linting
black src/
flake8 src/
mypy src/

Building

# Build package
python -m build

# Build Docker image
docker build -t voidlight-markitdown-mcp .

📊 Comparison with Microsoft MarkItDown

FeatureMicrosoft MarkItDownVoidLight MarkItDown MCP
Installationpip install markitdown-mcppip install voidlight-markitdown-mcp
SetupManual configurationAuto-setup with --setup
LanguagesEnglish focusKorean + Multilingual
Formats20+ formats50+ formats
Cloud ServicesBasicAzure + Google + OpenAI
PerformanceStandardOptimized + Caching
Docker
CLI ToolsBasicAdvanced management

🆘 Troubleshooting

Common Issues

Installation fails on macOS

# Install Python 3.11+ first
brew install python@3.11
pip3.11 install voidlight-markitdown-mcp

Missing dependencies

# Check status
voidlight-markitdown-mcp --status

# Install missing dependencies
voidlight-markitdown-mcp --install-deps all

Claude Desktop not detecting server

# Re-run setup
voidlight-markitdown-mcp --setup

# Check Claude Desktop config path
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json

📄 License

MIT License - see file for details.

🤝 Contributing

Contributions are welcome! Please see for guidelines.

🔗 Links


If this project helps you, please give it a star!