Web Search

Provides multi-engine web search and webpage content retrieval without API keys.

python

1

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "hughescuit-heventure-search-mcp": {
      "command": "heventure-search-mcp",
      "args": [],
      "env": {
        "PYTHONWARNINGS": "ignore"
      }
    }
  }
}

You can run a dedicated MCP server that searches the web using DuckDuckGo and Bing without needing API keys. It also lets you fetch the text content of a webpage. It runs asynchronously for high performance and can be used directly or integrated with an MCP client.

How to use

Set up the server locally and connect your MCP client to it. You can run the server as a native Python process or via a local MCP runner, then configure your MCP client to point at it so you can perform web searches across multiple engines or fetch webpage content.

How to install

Prerequisites: you need Python installed on your system. You may also use a package runner like uvx if you prefer running packages directly.

# Option A: install via PyPI and run the CLI
pip install heventure-search-mcp
heventure-search-mcp

# Option B: run with uvx (recommended for quick startup)
uvx heventure-search-mcp

# Option C: install from source via pip
pip install git+https://github.com/HughesCuit/heventure-search-mcp.git

# Then run one of the three start methods
heventure-search-mcp
# or
python -m server
# or
python -c "import server; import asyncio; asyncio.run(server.main())"

Configuration and runtime details

The server uses a standard browser user agent to avoid being blocked by sites.

User-Agent: 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'

Usage within an MCP client

To use this MCP server from your MCP client, reference the server as an MCP endpoint. The server supports multiple tools for searching and content extraction, described in the Tools section.

Notes on security and limits

No API keys are required, reducing credential exposure. Maintain a stable network connection and avoid excessive request rates. Respect terms of service for target sites and verify results independently.

Other practical tips

If you run the server locally, you can start it with a simple command and then point your MCP client at the local endpoint. You can also use a package runner like uvx for a quick startup in a development environment.

Example configurations and endpoints

The following examples illustrate how to configure the server for MCP usage. They show how to start the server locally and how to reference it from an MCP client.

Troubleshooting

If network requests fail, the server will retry automatically and provide detailed error logging. If parsing HTML content, it will gracefully degrade on errors and keep the response user-friendly.

Available tools

web_search

Performs web searches across configured engines (DuckDuckGo, Bing) and returns results with optional deduplication and sorting.

get_webpage_content

Fetches and returns the plain text content of a specified webpage URL.