web_search_claude_mcp_server_project

sukumar-govindraj/web_search_claude_mcp_server_project

3.2

If you are the rightful owner of web_search_claude_mcp_server_project and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

Claude-MCP is an AI assistant that integrates Anthropic's Claude LLM with a Model Context Protocol server for real-time web searches and tool-augmented reasoning.

Tools
1
Resources
0
Prompts
0

🧠 Claude-MCP Web-Integrated AI Assistant

Claude-MCP is an intelligent assistant that combines Anthropic’s Claude LLM with a lightweight Model Context Protocol (MCP) server to perform real-time web searches and tool-augmented reasoning.

It extracts meaningful answers from live web content using Claude’s tool-calling capabilities — ideal for building LLM agents that interact with APIs, services, or plugins.

🔍 Use Case

Ask anything like:

“What’s the latest news on GPT-5?”

Claude identifies this as a web query, calls the fetch_web_content tool, pulls search results via DuckDuckGo, and summarizes the answer.

Perfect for:

  • AI assistant prototyping
  • Custom tool-augmented LLM workflows
  • Demonstrating Claude's tool-use capabilities

📁 Project Structure

.
├── ask_claude.py              # CLI tool to ask Claude questions
├── claude_mcp_client.py       # LLM logic + tool call handling
├── mcp_integration.py         # MCP logic + DuckDuckGo search integration
├── mcp_server.py              # Flask server with /tool_call endpoint
├── requirements.txt           # Python dependencies

⚙️ Tech Stack

  • Claude v3 API (Anthropic)
  • Flask (MCP server)
  • DuckDuckGo (Search API)
  • Tool-calling Protocols
  • Python 3.8+

🚀 Quickstart Guide

2. Install Dependencies

pip install -r requirements.txt

3. Set Environment Variables

export CLAUDE_API_KEY=your_claude_api_key
export MCP_SERVER_URL=http://localhost:5001

4. Start the MCP Server

python mcp_server.py

5. Ask Claude a Question

python ask_claude.py "Tell me about quantum computing breakthroughs in 2024"

🧠 How It Works

  1. User asks a question via CLI.

  2. Claude interprets the query.

  3. If it detects a need for a web search, it:

    • Calls fetch_web_content
    • MCP server uses DuckDuckGo to fetch results
    • Returns results to Claude

  4. Claude summarizes it into a human-friendly answer.

🔌 Add Custom Tools

Define new tools in claude_mcp_client.py Handle logic in mcp_integration.py Expose via endpoint in mcp_server.py

Use cases:

  • Triggering APIs
  • Fetching database insights
  • Automating workflows

✅ Example Output

> python ask_claude.py "What is Claude 3 Opus?"
Searching for Claude 3 Opus
Answer: Claude 3 Opus is the most powerful LLM released by Anthropic...

📄 License

MIT License. Feel free to use, modify, and contribute!