中文 | English

🔍 MCP Web Search Server

Free forever. No API key required. A web search MCP server that works out of the box with Claude Desktop, Cursor, and any MCP-compatible AI tool.

pip install heventure-search-mcp

✨ Why?

Most MCP search servers require you to sign up for API keys (Bing, Google, SerpAPI...). This one works immediately — zero configuration, zero cost, zero sign-ups.

Feature	This Server	Others
No API key needed	✅	❌
DuckDuckGo (free)	✅	varies
Bing (free)	✅	❌
Google (free)	✅	❌
Optional SerpAPI/Tavily	✅	✅
Async + caching	✅	varies
Install in 10 seconds	✅	varies

🚀 Quick Start

Option 1: Claude Desktop / Cursor

Add to your MCP config:

{
  "mcpServers": {
    "web-search": {
      "command": "uvx",
      "args": ["heventure-search-mcp"]
    }
  }
}

Option 2: Command Line

pip install heventure-search-mcp
heventure-search-mcp

Option 3: Docker

docker run -p 8080:8080 heventure-search-mcp

🔧 Available Tools

`web_search`

Search the web with multiple engines simultaneously.

Parameter	Type	Default	Description
`query`	string	required	Search query
`max_results`	int	10	Number of results (1-20)
`search_engine`	string	`"both"`	`duckduckgo`, `bing`, `google`, or `both`

`get_page_content`

Extract readable text from any webpage.

Parameter	Type	Default	Description
`url`	string	required	Page URL to fetch

🔑 Optional: Enhanced Search

The free engines work great for most use cases. For higher quality results, you can optionally add paid API keys:

# SerpAPI — 100 free searches/month
export SERPAPI_KEY="your_key"

# Tavily — 1,000 free searches/month  
export TAVILY_API_KEY="your_key"

🏗️ Architecture

Engines: DuckDuckGo, Bing, Google, SerpAPI, Tavily
Caching: LRU cache with 300s TTL (100 entries max)
Protocol: MCP (Model Context Protocol)
Runtime: Python 3.10+ with asyncio

🤝 Contributing

Issues and Pull Requests are welcome! See CONTRIBUTING.md for guidelines.

📄 License

MIT License — use it however you want.

MCP Web Search Server

Documentation