Naver Search Docs MCP Server

API Gemini AI MCP(Model Context Protocol) FastAPI .

• API
• Gemini AI
• URL
• FastAPI REST API
• MCP

mcp_server/
 src/
    __init__.py
    main.py              # FastAPI 
    mcp_client.py        # MCP 
    mcp_server.py        # MCP  
    tools/               # MCP 
       __init__.py
       search_docs.py   #  
    utils/               #  
        __init__.py
        gemini.py        # Gemini AI 
        naver_search.py  #   
 tests/
    test_api.py          # API 
 docs/                    #  
    ai_deps              # Gemini  
    search_deps          #    
 requirements.txt         # Python 
 pyproject.toml          #  
 README.md               #  
 .gitignore              # Git  

1.

pip install -r requirements.txt

2.

API

src/utils/naver_search.py API :

NAVER_CLIENT_ID = 'your_client_id'
NAVER_CLIENT_SECRET = 'your_client_secret'

Google Cloud

src/utils/gemini.py Google Cloud :

os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "/path/to/your/credentials.json"

1: FastAPI ()

1. FastAPI

cd /Users/yanghyojun/Desktop/mcp_server
python -m src.main

uvicorn :

uvicorn src.main:app --reload --host 0.0.0.0 --port 8001

http://localhost:8001 .

2. API

** API **

curl -X POST "http://localhost:8001/search" \
  -H "Content-Type: application/json" \
  -d '{"query": "Python  "}'

---

{
  "result": "'Python  '    :\n\n1. Python  \n   URL: https://docs.python.org/ko/3/\n   : Python    ...\n\n",
  "query": "Python  "
}

**API **

• Swagger UI: http://localhost:8001/docs
• ReDoc: http://localhost:8001/redoc

---

• GET /: API
• GET /health:

3.

python tests/test_api.py

---

2:

MCP , FastAPI HTTP

1: MCP (8001 )

cd /Users/yanghyojun/Desktop/mcp_server
python -m src.main

2: HTTP

#  FastAPI  
import httpx
from fastapi import HTTPException

@app.post("/search-docs")
async def search_official_docs(query: str):
    """  """
    try:
        async with httpx.AsyncClient() as client:
            response = await client.post(
                "http://localhost:8001/search",
                json={"query": query},
                timeout=30.0
            )
            response.raise_for_status()
            return response.json()
    except httpx.RequestError as e:
        raise HTTPException(status_code=503, detail=f"MCP   : {str(e)}")

---

3: MCP

python -m src.mcp_server

Claude Desktop MCP :

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "naver-search-docs": {
      "command": "python",
      "args": ["-m", "src.mcp_server"],
      "cwd": "/Users/yanghyojun/Desktop/mcp_server"
    }
  }
}

MCP

1.

src/tools/ :

# src/tools/my_new_tool.py
from typing import Any, List
from mcp import types

class MyNewTool:
    """  """

    @staticmethod
    def get_tool_definition() -> types.Tool:
        """  """
        return types.Tool(
            name="my_new_tool",
            description=" ",
            inputSchema={
                "type": "object",
                "properties": {
                    "param": {
                        "type": "string",
                        "description": " "
                    }
                },
                "required": ["param"]
            }
        )

    @staticmethod
    async def execute(arguments: dict[str, Any]) -> List[types.TextContent]:
        """ """
        #   
        return [types.TextContent(
            type="text",
            text=""
        )]

2.

src/tools/__init__.py :

from .my_new_tool import MyNewTool

__all__ = [
    "SearchDocsTool",
    "MyNewTool",  # 
]

3. MCP

src/mcp_server.py TOOLS :

from .tools import SearchDocsTool, MyNewTool

TOOLS = [
    SearchDocsTool,
    MyNewTool,  # 
]

FastAPI

1. FastAPI MCP subprocess
2. MCP stdio MCP
3. REST API
4. MCP MCP
5. MCP API ( 30 )
6. Gemini AI
7. FastAPI JSON

MCP

1. Claude Desktop MCP MCP
3. API
4. Gemini AI
5. AI
6. , URL,

MIT