MCP Authentication Example
Demonstrates OAuth2/OIDC authentication for MCP servers using Asgardeo, with JWT validation and a sample weather tool to showcase secured API access.
README
MCP Auth Python Quick Start Guide - Asgardeo Integration
This project demonstrates how to create a secured MCP (Model Context Protocol) server in Python with FastMCP framework, using Asgardeo as the OAuth2/OIDC provider.
Prerequisites
- Python 3.12 or higher
- Asgardeo account and application setup
- pip (Python package installer)
Project Structure
├── main.py # Main FastMCP server application
├── jwt_validator.py # JWT validation module
├── .env # Environment variables configuration
├── README.md # This file
└── requirements.txt # Python dependencies
Installation
-
Clone or download this project
-
Create a virtual environment (recommended)
python -m venv .venv source .venv/bin/activate # On Windows: .venv\Scripts\activate -
Install required dependencies
pip install -r requirements.txt -
Configure environment variables
Create a
.envfile in the project root directory with your Asgardeo configuration:# Asgardeo OAuth2 Configuration AUTH_ISSUER=https://api.asgardeo.io/t/<your-tenant>/oauth2/token CLIENT_ID=<your-client-id> JWKS_URL=https://api.asgardeo.io/t/<your-tenant>/oauth2/jwksExample with actual values:
# Asgardeo OAuth2 Configuration AUTH_ISSUER=https://api.asgardeo.io/t/mycompany/oauth2/token CLIENT_ID=abc123xyz789_client_id_from_asgardeo JWKS_URL=https://api.asgardeo.io/t/mycompany/oauth2/jwks
Asgardeo Configuration
1. Create an Asgardeo Application
-
Login to your Asgardeo account.
-
Navigate to the Applications Tab and selct the MCP Client Application as shown in below image.

-
Add your application name and callback URL

2. Get Your Application Credentials
Once the application is created get both the Client ID and Tenant Name:
- Client ID: Found in the application's Protocol tab
- Tenant Name: Your organization's tenant name (visible in the URL)
3. Configure the Application
The application now uses environment variables for configuration. Make sure you have created the .env file as described in the Installation section above.
Replace the placeholders in your .env file:
- Replace
<your-tenant>with your actual Asgardeo tenant name - Replace
<your-client-id>with your actual OAuth2 client ID from Asgardeo
Security Note: Never commit your .env file to version control. Add .env to your .gitignore file to keep your credentials secure.
Running the Server
Development Mode
-
Start the server
python main.py -
Server will start on:
http://localhost:8000usingstreamable-httptransport
Test with MCP Inspector
Setup MCP Inspector
Run latest MCP Inspector UI with below command:
npx @modelcontextprotocol/inspector
For more information on how to run MCP Inspector, refer to the MCP Inspector documentation
Note: Make sure to get the Inspector release 0.16.3 or higher version.
Setup the MCP Inspector to connect with the server
- Once the mcp inspector got started, it will shows the URL where it is running, usually
http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=<Proxy_Token>. - Configure Inspector's callback URL in your Asgardeo application settings:
- Go to your Asgardeo application settings (You have created this in the previous steps).
- Navigate to Protocol tab
- Add the callback URL:
http://localhost:6274/oauth/callback(Port can be changed, make sure to match the port where MCP Inspector is running). - Update the application.
- Open the MCP Inspector in your browser using the URL provided by the MCP Inspector server, e.g.,
http://localhost:6274/?MCP_PROXY_AUTH_TOKEN=<Proxy_Token>. - Configure this sample mcp server as shown in below image:

- Click on the Connect button to establish a connection with the MCP server.
(Once you click on the connect button, it will redirect you to the Asgardeo login page, where you can login with your Asgardeo credentials.)

- Once connected, you can start testing the available get_weather tool by navigating the Tools tab in the MCP Inspector.
Available Tools
get_weather Tool
- Description: Get weather data for a city
- Parameters:
city(optional): City name (defaults to "London")
- Authentication: Required (valid JWT token)
- Returns: Weather information including temperature, condition, and humidity
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。