MCP Web Search Server

MCP Web Search Server is an MCP server that provides instant internet searching and webpage scraping capabilities directly to AI assistants without requiring external API keys. It connects language models running in environments like Claude Desktop or Cursor to multiple search providers, including DuckDuckGo, Bing, and Google, while also supporting optional keys for SerpAPI and Tavily. Software developers, researchers, and general AI users utilize this server to fetch up-to-date documentation, verify real-time facts, and pull readable text from websites during interactive chats. The server runs via Python 3.10+ using asynchronous requests, and it incorporates an in-memory LRU cache with a 300-second time-to-live to ensure fast repeat lookups and reduce network traffic. By offering native tools for web search and page content extraction, it enables conversational agents to browse the live internet out of the box with zero sign-ups or subscription costs.

Category: Web Search & Research

Tags: duckduckgo, information-retrieval, internet, web search

Visit MCP Web Search Server

How to install and configure MCP Web Search Server

Follow these steps to install and register MCP Web Search Server: 1. Ensure Python 3.10+ is installed on your system. 2. Open your client configuration file, such as the Claude Desktop or Cursor configuration. 3. Add the server under the mcpServers block using uvx: json { "mcpServers": { "web-search": { "command": "uvx", "args": ["heventure-search-mcp"] } } } 4. Alternatively, install via pip using pip install heventure-search-mcp and run heventure-search-mcp in your terminal, or launch using Docker with docker run -p 8080:8080 heventure-search-mcp. 5. Restart your MCP client to start using the search tools.

What you can do with MCP Web Search Server

• Querying recent software releases, frameworks, and technical documentation via DuckDuckGo without setting up API keys. * Extracting clean text from target documentation URLs using the get_page_content tool to provide full context. * Cross-checking factual information simultaneously across both Bing and Google search engines inside AI client interfaces. * Enhancing search depth and quality by optionally passing environment variables for SerpAPI or Tavily access. * Speeding up repetitive developer research queries through the built-in LRU cache with a 300-second TTL.

Key facts

• Free
• https://github.com/HughesCuit/heventure-search-mcp
• Web Search & Research
• duckduckgo, information-retrieval, internet, web search

Part of MCP Servers

How do I install MCP Web Search Server?

You can run the server directly using uvx by adding heventure-search-mcp to your MCP client configuration file. Alternatively, install it globally using pip install heventure-search-mcp and execute the command, or deploy it as a container using docker run -p 8080:8080 heventure-search-mcp.

What can MCP Web Search Server do?

The server exposes two main tools: web_search and get_page_content. The web_search tool queries DuckDuckGo, Bing, or Google without credentials, returning top results. The get_page_content tool fetches any given URL and extracts readable text for model processing.

Which MCP clients work with MCP Web Search Server?

It is designed to work out of the box with Claude Desktop, Cursor, and any other Model Context Protocol client supporting standard local process commands over stdio or Docker execution.

Does MCP Web Search Server require API keys?

No API keys are required for standard operation. The server searches DuckDuckGo, Bing, and Google freely. However, you can optionally configure SerpAPI or Tavily API keys as environment variables if you want higher quality or alternative search provider results.

Is MCP Web Search Server open source?

Yes, MCP Web Search Server is open source and distributed under the MIT License, allowing users to freely modify, deploy, and inspect the code via GitHub.