Claude-Deep-Research

Claude Deep Research is an MCP server that combines web and academic research capabilities to empower Claude and other MCP-compatible assistants. It integrates web search via DuckDuckGo and scholarly lookup through Semantic Scholar, enabling your AI to fetch current information, extract relevant content from pages, and present results with structured analysis and citations. The server is designed to work out of the box, but it can be customized via deep_research.py to adjust user agents, content size limits, and result counts. To use it, run the server using the defined command in your MCP client configuration and point Claude’s research tool to the provided endpoint. Once active, you can invoke the deep_research tool in conversations, providing a query and optional parameters for sources and the number of results. The tool will return a cohesive, multi-source research report with citations and suggested visualizations.

Prerequisites:

• Python 3.8 or higher
• pip (or uv package manager) installed

1. Install required Python packages
   • Run the quick install:

pip install mcp httpx beautifulsoup4

2. Clone the repository (if you haven’t already)

git clone https://github.com/yourusername/claude-deep-research.git

3. Prepare and run the server
   • Ensure you have a Python environment and the deep_research.py script in the repository.
   • Start the server (example):

python path/to/deep_research.py

4. Configure MCP client
   • In your MCP client (e.g., Claude Desktop), add a new MCP server entry using the provided mcp_config example and point to the python script path as shown in the configuration snippet.

Prerequisites overview:

• A functioning Python 3.8+ environment
• Internet access for web and academic searches
• Basic familiarity with editing configuration files for your MCP client

Tips and common issues:

• If you encounter connection failures, verify the path to deep_research.py is correct and that the Python environment is active.
• Some searches may timeout or return limited results; consider refining your query or increasing MAX_RESULTS in the config.
• The server relies on internet access for web and academic sources; ensure network access is not restricted.
• Content size may be capped; adjust MAX_CONTENT_SIZE and MAX_RESULTS in deep_research.py if you need more comprehensive outputs.
• For Claude Desktop integration, ensure the tool name matches (deep_research) and the path to the Python executable and script is correctly referenced in the Claude config.
• Citations are formatted in APA style by the tool; verify references if you require a different citation format.