Skip to content

MCP ToolHub 🚀

A learning-focused Model Context Protocol server that demonstrates every core MCP concept in one place.

  • 10 Tools

    From a simple echo to web search, screenshot capture, crypto prices, and OpenAI vector-store memory.

    Explore Tools

  • 3 Prompts

    Reusable message templates that guide the LLM — historical reports, topic analysis, weather explanations.

    Explore Prompts

  • 5 Resources

    URI-addressable read-only data — inventory items, prices, and weather statements.

    Explore Resources

  • Chat Prompt Recipes

    Copy-paste prompts to test every feature in Claude or any MCP-compatible client.

    Try Recipes

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI models interact with external tools, data sources, and services in a structured, safe way. Think of it as a USB-C port for AI — one universal interface that works across different models and clients.

graph LR
    A[Claude / AI Client] -->|MCP Protocol| B[MCP ToolHub Server]
    B --> C[🔧 Tools]
    B --> D[💬 Prompts]
    B --> E[📦 Resources]
    C --> F[APIs / Files / Screen]
    D --> G[Reusable Templates]
    E --> H[Read-only Data]

The Three Pillars

Concept What it does Examples here
Tools Callable functions the LLM can invoke echo, get_cryptocurrency_price, capture_screenshot, save_memory
Prompts Reusable message templates analyze_topic, historical_report, explain_weather_concept
Resources Read-only URI-addressable data inventory://overview, weather://{city}/statement

Quick Install

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-toolhub": {
      "command": "uvx",
      "args": ["--from", "/path/to/mcp-toolhub", "mcp-toolhub"],
      "env": {
        "OPENAI_API_KEY": "sk-...",
        "WEB_SEARCH_API_KEY": "pplx-..."
      }
    }
  }
}

unset VIRTUAL_ENV
uv run mcp dev mcpserver/myMcpServer.py

Full Quick Start