00_setup.md

Chapter 0 — Setup & Hosts

Install Everything

mkdir mcp-curriculum && cd mcp-curriculum
python -m venv .venv && source .venv/bin/activate

pip install fastmcp mcp anthropic httpx python-dotenv \
            starlette uvicorn pytest pytest-asyncio psutil

Create .env:

ANTHROPIC_API_KEY=sk-ant-your-key-here
GITHUB_TOKEN=ghp_your_token_here

---

Host 1 — Claude Desktop (Interactive)

Claude Desktop is a host: it spawns your server as a child process and connects to it. You test your server by chatting.

Install: Download from claude.ai/download

Config file location:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Add a server:

{
  "mcpServers": {
    "my-server": {
      "command": "python",
      "args": ["/absolute/path/to/server.py"],
      "env": {
        "GITHUB_TOKEN": "ghp_xxx"
      }
    }
  }
}

Restart Claude Desktop after every config change. Tools appear automatically in chat.

Subscription needed?

• Free tier: works, limited messages per day
• Pro ($20/mo): recommended for development — no message caps during builds

---

Host 2 — Claude API (Programmatic)

The Claude API lets you build your own client loop. You call Claude programmatically, handle tool_use responses, and route them to your MCP server or tool functions directly.

# verify_api.py
import anthropic
from dotenv import load_dotenv

load_dotenv()
client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=100,
    messages=[{"role": "user", "content": "Say hello in one word."}]
)
print(response.content[0].text)

Subscription needed?

• No subscription — pure pay-per-use
• Add credits at console.anthropic.com
• The entire curriculum costs roughly $5–10 in API calls

---

The Full Picture

You (developer)
    │
    ├── Claude Desktop ──────────► python server.py
    │   (chat UI, visual)              (your FastMCP server)
    │
    └── Python script ──────────► anthropic.Anthropic()
        (your custom client)           (Claude API)
             │
             └── also spawns ────► python server.py
                                       (same server!)

The same FastMCP server works with both hosts. You don't change the server — you change how you connect to it.

---

Hands-On: Verify Both Hosts Work

Step 1 — Verify Python and packages

python --version          # should be 3.10+
python -c "import fastmcp; print('FastMCP OK')"
python -c "import anthropic; print('Anthropic SDK OK')"

Step 2 — Verify Claude Desktop

Add this minimal server to your Claude Desktop config:

{
  "mcpServers": {
    "hello": {
      "command": "python",
      "args": ["/absolute/path/to/hello_server.py"]
    }
  }
}

# hello_server.py
from fastmcp import FastMCP

mcp = FastMCP("hello")

@mcp.tool()
def hello(name: str) -> str:
    """Say hello to someone."""
    return f"Hello, {name}! MCP is working."

if __name__ == "__main__":
    mcp.run()

Restart Claude Desktop. Ask: "Say hello to Alice using the hello tool."

Step 3 — Verify Claude API

# verify_tool_use.py
from anthropic import Anthropic
from dotenv import load_dotenv

load_dotenv()
client = Anthropic()

tools = [{
    "name": "hello",
    "description": "Say hello to someone.",
    "input_schema": {
        "type": "object",
        "properties": {"name": {"type": "string"}},
        "required": ["name"]
    }
}]

response = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=200,
    tools=tools,
    messages=[{"role": "user", "content": "Say hello to Bob."}]
)

for block in response.content:
    if block.type == "tool_use":
        print(f"Claude called: {block.name}({block.input})")
        # Execute the tool
        result = f"Hello, {block.input['name']}! MCP is working."
        print(f"Result: {result}")

Both should work before moving to Chapter 1.