Python MCP Weather Server with OAuth 2.1 Authentication
Provides weather information from the National Weather Service API with full MCP OAuth 2.1 compliance, including weather alerts and forecasts for US locations through secure Azure AD authentication.
README
Python MCP Weather Server with OAuth 2.1 Authentication
A production-ready Model Context Protocol (MCP) server built with FastAPI that provides weather information using the National Weather Service API. Features full MCP OAuth 2.1 compliance with PKCE, dynamic client registration, and Azure AD integration. Ready for deployment to Azure App Service with Azure Developer CLI (azd).
🌟 Features
- MCP OAuth 2.1 Specification Compliant: Complete implementation of MCP Authorization Specification (2025-03-26)
- PKCE Required: Secure authorization with Proof Key for Code Exchange (RFC 7636, S256 method)
- Dynamic Client Registration: Automatic client registration per RFC 7591
- Authorization Server Metadata: Discovery endpoint per RFC 8414
- Third-Party Authorization: Uses Azure AD as authorization server
- MCP Protocol Headers: Full support for
MCP-Protocol-Version: 2025-03-26 - JWT Token Management: Secure token-based authentication
- Weather Tools:
get_alerts: Get weather alerts for any US stateget_forecast: Get detailed weather forecast for any location
- Azure Ready: Pre-configured for Azure App Service deployment
- Web Test Interface: Built-in OAuth 2.1 flow testing
🔐 MCP Authorization Implementation
This server implements the complete MCP Authorization Specification (2025-03-26):
OAuth 2.1 Endpoints
GET /.well-known/oauth-authorization-server- Authorization server metadata (RFC 8414)POST /register- Dynamic client registration (RFC 7591)GET /authorize- Authorization endpoint with PKCE (RFC 7636)POST /token- Token endpoint for code exchange and refreshGET /auth/azure/callback- Third-party authorization callback
MCP Protocol Features
- ✅ Protocol Version Headers:
MCP-Protocol-Version: 2025-03-26 - ✅ PKCE Required: All clients must use S256 method
- ✅ Dynamic Registration: Automatic client onboarding
- ✅ JWT Authentication: Bearer token validation on MCP endpoints
- ✅ Proper Error Handling: 401/403/400 responses with details
- ✅ Azure AD Integration: Enterprise-grade authorization server
⚠️ Important: Complete OAuth setup required before use. See AUTH_SETUP.md for Azure AD configuration instructions.
💻 Local Development
Prerequisites
- Python 3.8+
- Azure account with completed OAuth setup (see AUTH_SETUP.md)
Setup & Run
-
Complete OAuth setup first: Follow the instructions in AUTH_SETUP.md to create your Azure App Registration.
-
Clone and install dependencies:
git clone <your-repo-url> cd remote-mcp-webapp-python-auth-oauth python -m venv venv .\venv\Scripts\Activate.ps1 # Windows # source venv/bin/activate # macOS/Linux pip install -r requirements.txt -
Configure environment variables:
cp .env.example .env # Edit .env with your Azure OAuth credentials from AUTH_SETUP.md -
Start the development server:
.\start_server.ps1 # Windows # or manually: uvicorn main:app --host 0.0.0.0 --port 8000 --reload -
Access the server:
- Server: http://localhost:8000/
- Health Check: http://localhost:8000/health
- OAuth 2.1 Test Interface: http://localhost:8000/mcp_oauth_test.html
- API Docs: http://localhost:8000/docs
🔌 Connect to the Local MCP Server
Authentication Required
Before connecting any MCP client, you must authenticate:
- Get JWT Token: Visit http://localhost:8000/mcp_oauth_test.html
- Complete OAuth Flow: Use the built-in OAuth 2.1 test interface
- Copy JWT Token: Use the token in your MCP client configuration
Using MCP Inspector
-
In a new terminal window, install and run MCP Inspector:
npx @modelcontextprotocol/inspector -
CTRL+click the URL displayed by the app (e.g. http://localhost:5173/#resources)
-
Configure authenticated connection:
- Set transport type to
HTTP - Set URL to:
http://localhost:8000/ - Add Authorization header:
Bearer <your-jwt-token>
💡 Getting your JWT token: Visit http://localhost:8000/mcp_oauth_test.html to complete the OAuth 2.1 flow and obtain your JWT token.
- Set transport type to
-
Test the connection: List Tools, click on a tool, and Run Tool
Configuration for MCP Clients
{
"mcpServers": {
"weather-mcp-server-local": {
"transport": {
"type": "http",
"url": "http://localhost:8000/",
"headers": {
"Authorization": "Bearer <your-jwt-token>"
}
},
"name": "Weather MCP Server (Local with Auth)",
"description": "Authenticated MCP Server with weather tools"
}
}
}
``` > 💡 **Replace `<your-jwt-token>`** with the actual JWT token obtained from the OAuth 2.1 flow at `/mcp_oauth_test.html`.
## 🚀 Quick Deploy to Azure
### Prerequisites
- [Azure CLI](https://docs.microsoft.com/en-us/cli/azure/install-azure-cli)
- [Azure Developer CLI (azd)](https://learn.microsoft.com/en-us/azure/developer/azure-developer-cli/install-azd)
- Active Azure subscription
- **Completed OAuth setup** (see [AUTH_SETUP.md](AUTH_SETUP.md))
### Deploy in 5 Commands
```bash
# 1. Login to Azure
azd auth login
# 2. Initialize the project
azd init
# 3. Set OAuth environment variables (from your AUTH_SETUP.md)
azd env set AZURE_CLIENT_ID "your-client-id"
azd env set AZURE_TENANT_ID "your-tenant-id"
azd env set AZURE_CLIENT_SECRET "your-client-secret"
azd env set JWT_SECRET_KEY "your-secure-jwt-secret"
# 4. Deploy to Azure (first time to get the URL)
azd up
# 5. Update environment with deployed URL and redeploy
azd env set BASE_URL "https://app-web-[unique-id].azurewebsites.net"
azd env set AZURE_REDIRECT_URI "https://app-web-[unique-id].azurewebsites.net/auth/azure/callback"
azd env set ENVIRONMENT "production"
azd up
Post-Deployment Setup
⚠️ Critical: After deployment, you must update your Azure App Registration:
- Note your deployed URL:
https://app-web-[unique-id].azurewebsites.net/ - Go to Azure Portal → Microsoft Entra ID → App registrations → Your App
- Click Authentication → Add redirect URI:
https://app-web-[unique-id].azurewebsites.net/auth/azure/callback - Click Save
💡 Note: The redirect URI must be
/auth/azure/callback(not/auth/callback) for the MCP OAuth 2.1 flow to work correctly.
Test Your Deployment
After deployment, your authenticated MCP server will be available at:
- OAuth 2.1 Test Interface:
https://<your-app>.azurewebsites.net/mcp_oauth_test.html - Health Check:
https://<your-app>.azurewebsites.net/health - MCP Capabilities:
https://<your-app>.azurewebsites.net/mcp/capabilities - API Docs:
https://<your-app>.azurewebsites.net/docs
🔌 Connect to the Remote MCP Server
Follow the same guidance as the local setup, but use your Azure App Service URL and ensure you have a valid JWT token from the deployed authentication endpoint.
Configuration for deployed server:
{
"mcpServers": {
"weather-mcp-server-azure": {
"transport": {
"type": "http",
"url": "https://<your-app>.azurewebsites.net/",
"headers": {
"Authorization": "Bearer <your-jwt-token>"
}
},
"name": "Weather MCP Server (Azure with Auth)",
"description": "Authenticated MCP Server hosted on Azure"
}
}
}
🧪 Testing
Interactive OAuth 2.1 Testing
- Local: Visit http://localhost:8000/mcp_oauth_test.html
- Azure: Visit
https://<your-app>.azurewebsites.net/mcp_oauth_test.html
The test interface provides:
- Complete OAuth 2.1 Flow: Dynamic client registration → Authorization → Token exchange
- PKCE Validation: Test the full Proof Key for Code Exchange flow
- MCP Endpoint Testing: Test authenticated weather tools
- JWT Token Display: View and validate your authentication tokens
- Client Callback Testing: Includes
/client-callbackendpoint for OAuth flow validation
OAuth Flow Architecture
The server implements a complete OAuth 2.1 flow:
- Client Registration: Dynamic client registration with auto-generated credentials
- Authorization: User redirected to Azure AD for authentication
- Azure Callback: Server receives Azure auth code at
/auth/azure/callback - Client Callback: Server redirects to client's callback (e.g.,
/client-callback) with authorization code - Token Exchange: Client exchanges authorization code for JWT access token
MCP Client Testing
Test with any MCP-compatible client using the authenticated endpoints and your JWT token.
🌦️ Data Source
This server uses the National Weather Service (NWS) API:
- Real-time weather alerts and warnings
- Detailed weather forecasts
- Official US government weather data
- No API key required
- High reliability and accuracy
🔒 Security Features
- ✅ OAuth 2.1 Compliance: Full MCP Authorization Specification implementation
- ✅ PKCE Required: S256 method for all authorization flows
- ✅ Dynamic Client Registration: Secure automatic client onboarding
- ✅ Azure AD Integration: Enterprise-grade authorization server
- ✅ JWT Token Security: Configurable expiration and secure validation
- ✅ Protocol Version Enforcement: MCP-Protocol-Version header validation
- ✅ Request Logging: Full audit trail with user identification
- ✅ CORS Protection: Proper cross-origin resource sharing policies
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。