duckduckgo-mcp-server

A Model Context Protocol (MCP) server that provides web search capabilities through DuckDuckGo, with additional features for content fetching and parsing. Implemented in both TypeScript and Python.

Source: https://github.com/nickclyde/duckduckgo-mcp-server

Features

• Web Search: Enables DuckDuckGo web search with advanced rate limiting and result formatting.
• Content Fetching: Retrieves and parses webpage content with intelligent text extraction.
• Rate Limiting:
  • Search: 30 requests per minute
  • Content fetching: 20 requests per minute
  • Automatic queue management and wait times
• Error Handling: Comprehensive error catching and logging, with graceful degradation on rate limits or timeouts.
• LLM-Friendly Output: Results formatted for optimal consumption by large language models (LLMs).
• Result Processing:
  • Removes ads and irrelevant content
  • Cleans up DuckDuckGo redirect URLs
  • Truncates long content where needed
• Tools:
  • search(query: str, max_results: int = 10): Performs a DuckDuckGo search and returns formatted results (titles, URLs, snippets).
  • fetch_content(url: str): Fetches and cleans text content from a webpage.
• Installation options:
  • Via Smithery (npx -y @smithery/cli install @nickclyde/duckduckgo-mcp-server --client claude)
  • Via PyPI using uv pip install duckduckgo-mcp-server

Pricing

• Open source, licensed under the MIT License. No cost to use.

Category

• data-access-integration-mcp-servers

Tags

• mcp
• duckduckgo
• web-search
• privacy