duckduckgo

The DuckDuckGo MCP Server implements the MCP web search toolset using DuckDuckGo as the backend. It exposes a single tool named search that accepts a query string and returns an array of web results. To use it, deploy the MCP server in your environment (locally via Node.js, or via Docker/UVX per the provided examples), then register the duckduckgo MCP server in your Claude Desktop configuration. Once configured, you can invoke the search tool by sending a query; the server will process the request, perform a DuckDuckGo web search, and return structured results suitable for summarization or further processing.

Prerequisites:

• Node.js (recommended >14) and npm
• Optional: Docker if you prefer running the Docker image

Install locally:

1. Ensure Node.js and npm are installed
2. Clone or download the MCP server repository
3. Install dependencies (if a package.json exists):

npm install

4. Run the server (assuming server.js is the entrypoint):

node server.js

Run with Docker (alternative):

docker run -i --rm ashdev/duckduckgo-mcp-server

If you want to configure Claude Desktop to use UVX or Docker, you can follow the examples in the README to point to the appropriate command and entrypoint.

Notes and tips:

• The server provides a single tool: search. Ensure the input uses a plain string query. The response is an array of results from DuckDuckGo.
• If deploying in a restricted environment, consider setting a request timeout and handling network errors gracefully in your integration layer.
• When running via Docker, ensure network access to the container and that any required ports are exposed if needed by your setup.
• There is no required environment variable documented for this server in the README; if you extend functionality, you can add env vars such as API or rate-limit settings and describe them in the environment configuration.