FILE CONVERTER MCP

🚀 Pandoc Multi-Server Control Protocol (MCP) Server

Pandoc-MCP Server is a Pandoc-based tool that supports the conversion and processing of various file formats. This document details how to build and run the server and integrate it with MCP clients such as LangChain.

🚀 Quick Start

Step 1: Build the Docker Image

docker build -t your_username/pandoc-mcp-server .

Step 2: Run the Container

docker run -it --rm -v $(pwd):/data your_username/pandoc-mcp-server:latest

✨ Features

• Multi-format Conversion: Supports the mutual conversion of files in formats such as Markdown, Word, and PDF.
• Remote Control: Provides services via the HTTP protocol and supports remote calls for file processing functions.
• Docker Support: Offers a Docker image building solution for easy and rapid deployment and operation.
• Document Management: Supports batch processing, version control, and document storage.

📦 Installation

Server-side Installation

Option 1: Docker-based Installation (Recommended)

1. Pull the Image

   docker pull your_username/pandoc-mcp-server:latest
   

2. Start the Container

   docker run -it --rm -v $(pwd):/data your_username/pandoc-mcp-server:latest
   

3. Access the Service: After the server starts, you can access the service via http://localhost:8000/sse.

Option 2: Local Installation (Requires Manual Configuration)

1. Install Dependencies

   • Install Python 3.10 or a higher version.
   • Install the Pandoc toolchain (pandoc --version).
   • Install the LaTeX environment (pdflatex --version).

2. Clone the Repository

   git clone https://github.com/your_username/pandoc-mcp-server.git
   cd pandoc-mcp-server
   

3. Create a Virtual Environment (Recommended)

   python -m venv venv
   source venv/bin/activate  # Linux/macOS
   # Or: venv\Scripts\activate  # Windows
   

4. Install Python Packages

   pip install -r requirements.txt
   

5. Start the Server

   python pandoc_mcp_server.py
   

💻 Usage Examples

Client Integration Example (Taking LangChain as an Example)

from langchain.agents import MultiServerMCPClient

# Initialize the client
client = MultiServerMCPClient(
    server_url="http://localhost:8000/sse",
    token="your_token"
)

# Invoke the file conversion service
response = client.invoke("Convert input.md to PDF")
print(response)

📚 Documentation

1. File Format Conversion

• Supported Formats: Markdown, .docx, PDF, etc.
• Usage Commands

  # Convert Markdown to PDF
  ./pandoc-mcp convert input.md output.pdf --format=pdf
  
  # Batch process files
  ./pandoc-mcp batch process *.md output/
  

2. Document Management

• Version Control: Supports the management of historical versions of files.
• Storage Location: By default, files are stored in the /data/documents directory.

3. Advanced Features

• Custom Templates: You can use custom Pandoc templates by configuring the templates directory.
• Logging: The server supports detailed log output for easy debugging and monitoring.

Example Scenarios

Scenario 1: Online Document Conversion Service

# Start the service and bind to port 8000
docker run -it --rm -p 8000:8000 your_username/pandoc-mcp-server:latest

Scenario 2: Batch Process Word Files

# Convert all .docx files in the current directory to PDF
./pandoc-mcp batch convert *.docx output/pdf/

Project Structure

.
├── pandoc_mcp_server.py    # Main program entry
├── requirements.txt        # Python dependency management
├── Dockerfile              # Docker image building file
└── templates/             # Pandoc template directory

Frequently Asked Questions

1. How to Get Help?

   • Check pandoc-mcp --help.
   • Join the project community discussion group or submit an ISSUE.

2. How to Handle Errors?

   • Check the log output (docker logs).
   • Ensure that all dependencies are correctly installed.

🤝 Project Contribution

Welcome to fork and submit PRs!

• Submit feature enhancements.
• Fix known issues.
• Add support for more file formats.

Through the above steps, you can easily set up and use the Pandoc-MCP Server to achieve efficient document management and conversion.