Pinkpixel Dev Web Scout MCP

What pinkpixel-dev-web-scout-mcp MCP server does

The pinkpixel-dev-web-scout-mcp MCP server gives MCP clients two web-research capabilities: DuckDuckGo search and webpage content extraction. Search returns structured findings containing titles, URLs, and snippets. Content extraction removes page elements such as scripts, styles, and navigation so an agent can work with the main readable text.

The server is suited to agents that need to discover webpages, inspect articles or documentation, or process several known URLs without handling HTML cleanup themselves. It does not present a general browser automation interface; the documented functionality is limited to search and text extraction.

How it works

The DuckDuckGoWebSearch tool accepts a required string query and an optional maxResults number. If the limit is omitted, the default is 10 results. The search provider is DuckDuckGo.

The UrlContentExtractor tool accepts either one URL string or an array of URL strings. When multiple URLs are supplied, the server can extract them in parallel. Its output is the extracted text content for the requested pages.

The implementation includes request throttling intended to reduce API blocks, memory-management measures for larger workloads, and error handling for failed operations. The README does not specify the exact rate limits, extraction library, supported page types, or error response format.

Setup and configuration

Use Node.js 18 or newer with npm or yarn. Install the package globally with npm install -g @pinkpixel/web-scout-mcp, install it locally with npm install @pinkpixel/web-scout-mcp, or run it through npx in an MCP client configuration.

For a stdio client, configure the server with the npx command and the -y and @pinkpixel/web-scout-mcp@latest arguments. The README specifically shows Claude Desktop and Cursor as configuration examples. A globally installed package can also be started with web-scout-mcp.

Set WEB_SCOUT_DISABLE_AUTOSTART=1 when embedding the package and invoking createServer() yourself. The published entrypoint otherwise starts the stdio transport automatically.

Tools and capabilities

• Search DuckDuckGo with a query and optional result limit.
• Return result titles, URLs, and snippets.
• Extract cleaned text from one webpage.
• Extract text from multiple webpages in one request.
• Process multiple URL requests in parallel.
• Apply request throttling and memory-conscious processing.

Limitations and notes

The pinkpixel-dev-web-scout-mcp MCP server requires network access to search DuckDuckGo and fetch webpages. The provided material does not document authentication settings, proxy configuration, JavaScript-rendered page support, domain allowlists, or a separate HTTP transport.

The repository materials contain conflicting license indicators: a badge identifies MIT, while the license section describes Apache 2.0. Verify the repository license file before redistributing the package.