Tutorial: Create and connect a client application to your Azure IoT Central application (Python)

This article applies to solution builders and device developers.

This tutorial shows you how, as a device developer, to connect a Python client application to your Azure IoT Central application. The Python application simulates the behavior of an environmental sensor device. You use a sample device capability model to create a device template in IoT Central. You add views to the device template to enable an operator to interact with a device.

In this tutorial, you learn how to:

  • Import a device capability model to create a device template.
  • Add default and custom views to a device template.
  • Publish a device template and add a real device to your IoT Central application.
  • Create and run the Python device code and see it connect to your IoT Central application.
  • View the simulated telemetry sent from the device.
  • Use a view to manage device properties.
  • Call synchronous and asynchronous commands to control the device.


To complete the steps in this article, you need the following:

  • An Azure IoT Central application created using the Custom application template. For more information, see the create an application quickstart.
  • A development machine with Python version 3.7 or later installed. You can run python3 --version at the command line to check your version. Python is available for a wide variety of operating systems. The instructions in this tutorial assume you're running the python3 command at the Windows command prompt.

Create a device template

Create a folder called environmental-sensor on your local machine.

Download the Environmental sensor capability model JSON file and save it in the environmental-sensor folder.

Use a text editor to replace the two instances of {YOUR_COMPANY_NAME_HERE} with your company name in the EnvironmentalSensorInline.capabilitymodel.json file you downloaded. Use only the characters a-z, A-Z, 0-9, and underscore.

In your Azure IoT Central application, create a device template called Environmental sensor by importing the EnvironmentalSensorInline.capabilitymodel.json device capability model file:

Device template with imported device capability model

The device capability model includes two interfaces: the standard Device Information interface and the custom Environmental Sensor interface. The Environmental Sensor interface defines the following capabilities:

Type Display Name Description
Property Device State The state of the device. Two states online/offline are available.
Property (writeable) Customer Name The name of the customer currently operating the device.
Property (writeable) Brightness Level The brightness level for the light on the device. Can be specified as 1 (high), 2 (medium), 3 (low).
Telemetry Temperature Current temperature detected by the device.
Telemetry Humidity Current humidity detected by the device.
Command blink Begin blinking the LED on the device for given time interval.
Command turnon Turn on the LED on the device.
Command turnoff Turn off the LED on the device.
Command rundiagnostics This asynchronous command starts a diagnostics run on the device.

To customize how the Device State property displays in your IoT Central application, select Customize in the device template. Expand the Device State entry, enter Online as the True name and Offline as the False name. Then save the changes:

Customize the device template

Create views

Views let you interact with devices connected to your IoT Central application. For example, you can have views that display telemetry, views that display properties, and views that let you edit writeable and cloud properties. Views are part of a device template.

To add some default views to your Environmental sensor device template, navigate to your device template, select Views, and select the Generate Default views tile. Make sure Overview and About are On, and then select Generate default dashboard view(s). You now have two default views defined in your template.

The Environmental Sensor interface includes two writeable properties - Customer Name and Brightness Level. To create a view, you can use to edit these properties:

  1. Select Views and then select the Editing device and cloud data tile.

  2. Enter Properties as the form name.

  3. Select the Brightness Level and Customer Name properties. Then select Add section.

  4. Save your changes.

Add a view to enable property editing

Publish the template

Before you can add a device that uses the Environmental sensor device template, you must publish it.

In the device template, select Publish. On the Publish this device template to the application panel, select Publish.

To check that the template is ready to use, navigate to the Devices page in your IoT Central application. The Devices section shows a list of the published devices in the application:

Published templates on the devices page

Add a real device

In your Azure IoT Central application, add a real device to the device template you created in the previous section:

  1. On the Devices page, select the Environmental sensor device template.


    Be sure to select the template to use before you select + New, otherwise you'll create an unassociated device.

  2. Select + New.

  3. Make sure that Simulated is Off. Then select Create.

Click on the device name, and then select Connect. Make a note of the device connection information on the Device Connection page - ID scope, Device ID, and Primary key. You need these values when you create your device code:

Device connection information

Create a Python application

The following steps show you how to create a Python client application that connects to the real device you added to the application. This Python application simulates the behavior of a real device.

  1. In your command-line environment, navigate to the environmental-sensor folder you created previously.

  2. To install the required libraries, run the following commands:

    pip install azure-iot-device
  3. Create a file called environmental_sensor.py in the environmental-sensor folder.

  4. Add the following import statements at the start of the environmental_sensor.py file:

    import asyncio
    import os
    import json
    import datetime
    import random
    from azure.iot.device.aio import ProvisioningDeviceClient
    from azure.iot.device.aio import IoTHubDeviceClient
    from azure.iot.device import MethodResponse
    from azure.iot.device import Message
  5. Add the following asynchronous main function and variable declarations to the file:

    async def main():
      # In a production environment, don't store
      # connection information in the code.
      provisioning_host = 'global.azure-devices-provisioning.net'
      id_scope = '{your Scope ID}'
      registration_id = '{your Device ID}'
      symmetric_key = '{your Primary Key}'
      delay = 2
      # All the remaining code is nested within this main function
    if __name__ == '__main__':

    Update the placeholders {your Scope ID}, {your Device ID}, and {your Primary Key} with the values you made a note of previously. In a real application, don't hard code this information in the application.

    All the following function definitions and code are nested within the main function.

  6. Add the following two functions inside the main function to register the device and connect it to your IoT Central application. Registration uses the Azure Device Provisioning Service:

      async def register_device():
        provisioning_device_client = ProvisioningDeviceClient.create_from_symmetric_key(
        registration_result = await provisioning_device_client.register()
        print(f'Registration result: {registration_result.status}')
        return registration_result
      async def connect_device():
        device_client = None
          registration_result = await register_device()
          if registration_result.status == 'assigned':
            device_client = IoTHubDeviceClient.create_from_symmetric_key(
            # Connect the client.
            await device_client.connect()
            print('Device connected successfully')
          return device_client
  7. Add the following function inside the main function to send telemetry to your IoT Central application:

      async def send_telemetry():
        print(f'Sending telemetry from the provisioned device every {delay} seconds')
        while True:
          temp = random.randrange(1, 75)
          humid = random.randrange(30, 99)
          payload = json.dumps({'temp': temp, 'humid': humid})
          msg = Message(payload)
          await device_client.send_message(msg, )
          print(f'Sent message: {msg}')
          await asyncio.sleep(delay)

    The names of the telemetry items (temp and humid) must match the names used in the device template.

  8. Add the following functions inside the main function to handle commands called from your IoT Central application:

      async def blink_command(request):
        print('Received synchronous call to blink')
        response = MethodResponse.create_from_method_request(
          request, status = 200, payload = {'description': f'Blinking LED every {request.payload} seconds'}
        await device_client.send_method_response(response)  # send response
        print(f'Blinking LED every {request.payload} seconds')
      async def diagnostics_command(request):
        print('Starting asynchronous diagnostics run...')
        response = MethodResponse.create_from_method_request(
          request, status = 202
        await device_client.send_method_response(response)  # send response
        print('Generating diagnostics...')
        await asyncio.sleep(2)
        print('Generating diagnostics...')
        await asyncio.sleep(2)
        print('Generating diagnostics...')
        await asyncio.sleep(2)
        print('Sending property update to confirm command completion')
        await device_client.patch_twin_reported_properties({'rundiagnostics': {'value': f'Diagnostics run complete at {datetime.datetime.today()}.'}})
      async def turnon_command(request):
        print('Turning on the LED')
        response = MethodResponse.create_from_method_request(
          request, status = 200
        await device_client.send_method_response(response)  # send response
      async def turnoff_command(request):
        print('Turning off the LED')
        response = MethodResponse.create_from_method_request(
          request, status = 200
        await device_client.send_method_response(response)  # send response
      commands = {
        'blink': blink_command,
        'rundiagnostics': diagnostics_command,
        'turnon': turnon_command,
        'turnoff': turnoff_command,
      # Define behavior for handling commands
      async def command_listener():
        while True:
          method_request = await device_client.receive_method_request()  # Wait for commands
          await commands[method_request.name](method_request)

    The names of the commands (blink, turnon, turnoff, and rundiagnostics) must match the names used in the device template.

    Currently, IoT Central doesn't use the response schema defined in the device capability model. For a synchronous command, the response payload can be any valid JSON. For an asynchronous command, the device should return a 202 response immediately, followed by reported property update when the work is finished. The format of the reported property update is:

      [command name] : {
        value: 'response message'

    An operator can view the response payload in the command history.

  9. Add the following functions inside the main function to handle property updates sent from your IoT Central application:

        async def name_setting(value, version):
          await asyncio.sleep(1)
          print(f'Setting name value {value} - {version}')
          await device_client.patch_twin_reported_properties({'name' : {'value': value['value'], 'status': 'completed', 'desiredVersion': version}})
        async def brightness_setting(value, version):
          await asyncio.sleep(5)
          print(f'Setting brightness value {value} - {version}')
          await device_client.patch_twin_reported_properties({'brightness' : {'value': value['value'], 'status': 'completed', 'desiredVersion': version}})
        settings = {
          'name': name_setting,
          'brightness': brightness_setting
        # define behavior for receiving a twin patch
        async def twin_patch_listener():
          while True:
            patch = await device_client.receive_twin_desired_properties_patch() # blocking
            to_update = patch.keys() & settings.keys()
            await asyncio.gather(
              *[settings[setting](patch[setting], patch['$version']) for setting in to_update]

    When the operator sets a writeable property in the IoT Central application, the application uses a device twin desired property to send the value to the device. The device then responds using a device twin reported property. When IoT Central receives the reported property value, it updates the property view with a status of synced.

    The names of the properties (name and brightness) must match the names used in the device template.

  10. Add the following functions inside the main function to control the application:

      # Define behavior for halting the application
      def stdin_listener():
        while True:
          selection = input('Press Q to quit\n')
          if selection == 'Q' or selection == 'q':
      device_client = await connect_device()
      if device_client is not None and device_client.connected:
        print('Send reported properties on startup')
        await device_client.patch_twin_reported_properties({'state': 'true'})
        tasks = asyncio.gather(
        # Run the stdin listener in the event loop
        loop = asyncio.get_running_loop()
        user_finished = loop.run_in_executor(None, stdin_listener)
        # Wait for user to indicate they are done listening for method calls
        await user_finished
        # Cancel tasks
        tasks.add_done_callback(lambda r: r.exception())
        await device_client.disconnect()
        print('Device could not connect')
  11. Save the the environmental_sensor.py file.

Run your Python application

To start the device client application, run the following command in your command-line environment:

python3 environmental_sensor.py

You can see the device connects to your Azure IoT Central application and starts sending telemetry:

Run the client application

As an operator in your Azure IoT Central application, you can:

  • View the telemetry sent by the device on the Overview page:

    View telemetry

  • View the device properties on the About page:

    View properties

  • Update writeable property values on the Properties page:

    Update properties

  • Call the commands from the Commands page:

    Call blink command

    View command history

You can see how the device responds to commands and property updates:

Observe the client application

Next steps

As a device developer, now that you've learned the basics of how to create a device using Python, some suggested next steps are to:

If you'd prefer to continue through the set of IoT Central tutorials and learn more about building an IoT Central solution, see: