mersdev/duckduckgo-mcp

DuckDuckGo Search MCP

A Python-based MCP (Model Context Protocol) server that provides specialized search functionality using DuckDuckGo's search engine. This implementation offers three distinct search tools:

• GitHub Code Search
• Malaysia News Search (NST)
• Azure Documentation Search

Features

1. GitHub Code Search

Search for code examples and repositories on GitHub related to specific topics.

• Automatically filters results to GitHub domain
• Returns top 5 relevant results with titles, links, and descriptions

2. Malaysia News Search

Search for local news articles from New Straits Times (NST) Malaysia.

• Specifically targets nst.com.my domain
• Includes results from the past year
• Returns top 5 news articles with titles, links, and descriptions

3. Azure Documentation Search

Search for official Azure documentation on Microsoft Learn.

• Focuses on Azure-specific documentation
• Returns top 5 relevant documentation pages with titles, links, and descriptions

Installation

1. Ensure you have Python 3.11 or later installed

2. Clone the repository:

git clone <repository-url>
cd duckduckgo-mcp

3. Install dependencies using pip:

pip install -r requirements.txt

Required dependencies:

• duckduckgo-search>=4.1.1
• mcp[cli]>=1.6.0

Usage

1. Start the MCP server:

python search.py

2. The server will initialize with three available search tools:

   • search_github_code(topic: str)
   • search_malaysia_news(topic: str)
   • search_azure_docs(topic: str)

3. Each search function accepts a topic string and returns formatted results including titles, links, and descriptions.

Example Usage

# Search for Python asyncio code examples on GitHub
search_github_code("Python asyncio")

# Search for technology news in Malaysia
search_malaysia_news("technology startup")

# Search for Azure Kubernetes Service documentation
search_azure_docs("AKS cluster creation")

Project Setup

Using uv Package Manager (Recommended)

1. If you haven't installed uv yet, install it first:

pip install uv

2. Create a new MCP project:

uv init duckduckgo-mcp
cd duckduckgo-mcp

3. Add MCP to your project dependencies:

uv add "mcp[cli]"

4. Run the MCP server:

uv run mcp

Using pip (Alternative)

Alternatively, you can use pip to install dependencies:

pip install "mcp[cli]"

Development

Basic Development

The project uses FastMCP for creating the MCP server and DuckDuckGo Search API for performing searches. Each search function is decorated with @mcp.tool() to make it available as an MCP tool.

Development Mode

The fastest way to test and debug your server is with the MCP Inspector:

# Run server in development mode
mcp dev search.py --with duckduckgo-search

Development mode provides:

• Interactive debugging interface
• Hot reload for code changes
• Real-time tool testing
• Dependency management
• Local code mounting for development

Cline Extension Integration

To integrate the DuckDuckGo Search MCP with CLine extension, add the following configuration to your CLine settings:

"duckduckgo-search": {
    "autoApprove": [],
    "disabled": false,
    "timeout": 60,
    "command": "uv",
    "args": [
        "run",
        "--with",
        "mcp[cli]",
        "--with",
        "duckduckgo-search",
        "mcp",
        "run",
        "/path/to/example/search.py"
    ],
    "transportType": "stdio"
}

This configuration:

• Uses uv package manager to run the MCP server
• Sets a 60-second timeout for operations
• Automatically loads required dependencies (mcp[cli] and duckduckgo-search)
• Configures stdio transport for communication
• Specifies the path to your search.py implementation