mcp-search

The mcp-search server exposes a modular registry of 27 tools for web search, content extraction, and data processing, designed to be used by language models and automation pipelines without requiring API keys. It runs as a Python package, serving as a centralized hub from which you can discover, invoke, and compose the available tools (such as web search viaDuckDuckGo, content extraction from web pages, and academic content helpers). The server dynamically loads tools through a modular registry, enabling discovery commands like search_tools, list_tool_categories, and get_tool_info to help you understand what’s available and how to use each tool. Once started, you can interact with the server by querying specific tools or by using the meta-tools to explore and select the right tool for a task.

Prerequisites:

• Python 3.10+ (recommended)
• pip (comes with Python)

Install from PyPI (recommended):

pip install mcp-search-server

Install from source:

git clone https://github.com/KazKozDev/mcp-search-server.git
cd mcp-search-server
pip install -e .

Run the server (example using the module entry point):

python -m mcp_search_server

Optional: If you want Claude or another client to reference the server directly, configure the appropriate client integrations to point at the running server address (e.g., http://localhost:8000 by default, depending on the package's defaults).

Notes and tips:

• This server requires no API keys and relies on public data sources configured by the integrated tools.
• The tool registry is modular; new tools can be added by updating the package or registry configuration.
• If you upgrade from an older version, re-check tool metadata with get_tool_info to ensure parameters and usage examples are up to date.
• If the server cannot load a tool, verify dependencies are installed (e.g., requests, beautifulsoup4, pypdf2, pdfplumber) as needed by specific tools.
• When deploying in production, consider running behind a reverse proxy and enabling appropriate logging and rate limiting as supported by the server framework.
• Review the available tools via the meta-tools (search_tools, list_tool_categories) to plan task workflows and tool-chains.