Provides tools for searching the web via DuckDuckGo (search) and fetching webpage content with LLM-based relevance extraction (fetch_content). No API key required.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@DuckDuckGo MCP Server search for latest developments in artificial intelligence"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A lightweight Model Context Protocol (MCP) server that gives AI agents the ability to search the web via DuckDuckGo and fetch content from any URL — no API key required.
Web search — query DuckDuckGo and receive ranked results with titles, URLs, and snippets
Page fetching — retrieve and parse the readable text content of any webpage
LLM-powered relevance extraction — fetch_content routes raw page text through a local/remote LLM and returns only the passages relevant to the query
Rate limiting — built-in sliding-window rate limiter protects against throttling
Stateless HTTP transport — compatible with any MCP client that speaks Streamable HTTP
Health endpoint — /actuator/health for container orchestration probes
Docker-ready — minimal Alpine-based image included
Related MCP server: search-mcp
Tool | Description | Input | Output |
| Search DuckDuckGo and return the top N results |
| Ranked list of results (title, URL, snippet) |
| Fetch a webpage and return only the content relevant to the query |
| Passages from the page relevant to the query, extracted by an LLM |
searchQueries the DuckDuckGo HTML endpoint (no JavaScript required), parses the results, and returns them in a structured natural-language format suited for LLM consumption. Ad results are automatically filtered out.
Found 5 search results:
1. Example Title
URL: https://example.com
Summary: A short snippet describing the page content.
...fetch_contentFetches a URL with httpx, strips navigation, headers, footers, scripts, and style elements, then passes the cleaned text together with the original query to a local/remote LLM (ContentExtractor). The LLM returns only the passages that are directly relevant to the query, preserving the original wording without summarising or paraphrasing. If no relevant content is found, the tool returns "The fetched page does not contain content relevant to the query."
The LLM endpoint must be OpenAI-compatible and is configured via the LLM_BASE_URL and LLM_MODEL environment variables (see Configuration).
Python >= 3.10
git clone https://github.com/your-org/duckduckgo-mcp-server.git
cd duckduckgo-mcp-server
# Create and activate a virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate
pip install -e .Copy the example environment file and adjust as needed:
cp .env.example .envVariable | Default | Description |
|
| Port the HTTP server listens on |
|
| Maximum number of search results returned per query |
| (required) | Base URL of the OpenAI-compatible LLM endpoint used by |
|
| Model name passed to the LLM endpoint |
python server.pyThe server starts on http://0.0.0.0:<PORT> and exposes:
Path | Description |
| MCP Streamable HTTP endpoint |
| Health check — returns |
docker build -t duckduckgo-mcp-server .docker run -p 8080:8080 duckduckgo-mcp-serverdocker run -p 9000:9000 \
-e PORT=9000 \
-e MAX_RESULTS=10 \
duckduckgo-mcp-serverConfigure your MCP client to connect to the server's Streamable HTTP endpoint. Below are examples for common clients.
Add the following to your claude_desktop_config.json:
{
"mcpServers": {
"duckduckgo": {
"url": "http://localhost:8080/mcp"
}
}
}Transport: Streamable HTTP
URL: http://localhost:8080/mcpBoth tools enforce an in-process sliding-window rate limiter to avoid triggering DuckDuckGo's bot-detection or overwhelming remote servers:
Component | Limit |
| 30 requests / minute |
| 20 requests / minute |
When the limit is reached the request waits automatically — no error is raised.
The search parsing and webpage fetching logic is inspired by nickclyde/duckduckgo-mcp-server.
This server cannot be deployed
LLM-ready web search + instant answers + URL-to-clean-text fetch for agents and RAG.
Web search, browser automation, scraping, crawling and CAPTCHA solving for AI agents.
Web search and page-reading for AI agents. One-click OAuth connect, or a Caesar API key.
Web search, fetch, extract, and research for AI agents. Markdown output + AI-synthesized answers.
