Azure Weather SSE

You run a remote MCP server that streams updates using SSE and protects access with an API key. This server is built with Python and can be run locally or deployed to Azure Container Apps, allowing you to connect from MCP clients to fetch data and trigger actions in real time.

How to use

Connect your MCP client to the server using the remote URL that streams events via SSE. Provide your API key when the client initiates the session or configures the connection so that requests are authenticated. You can also run the server locally and connect to http://localhost:8000/sse if you are testing in a local development environment. Use the local URL with your MCP client to receive a continuous stream of updates and to perform actions exposed by the server.

How to install

Prerequisites before you start:

• Python 3.11 or later
• uv command-line tool (provided by the uv package)

Install and run the server locally on your machine using the following commands in sequence.

uv venv
uv sync

# linux/macOS
export API_KEYS=<AN_API_KEY>
# Windows
set API_KEYS=<AN_API_KEY>

uv run fastapi dev main.py

Additional sections

Deployments and configuration for the MCP server are shown here for a practical workflow. The following demonstrates how to deploy to Azure Container Apps and how to prepare a local development environment for testing.

Azure Container Apps deployment example shows how to publish the server so your MCP clients can connect from anywhere. Use the following command to deploy, passing your resource group, environment, region, and a key for authentication.

az containerapp up -g <RESOURCE_GROUP_NAME> -n weather-mcp --environment mcp -l westus --env-vars API_KEYS=<AN_API_KEY> --source .

If the deployment is successful, the Azure CLI returns the URL of the app. You can use this URL to connect to the server from your MCP client. If the deployment fails, update the CLI and extension and retry with these commands.

az upgrade
az extension add -n containerapp --upgrade