AdvantageAir2MQTT
A new Daikin reverse cycle central air conditioning system, to replace the old gas unit, came with a zoning duct system. The zoning duct work involves motorized vanes which direct air flow. All of this is under the control of a software system called MyAir (https://www.advantageair.com.au/myair/). The challenge is to integrate the system with the home automation (OpenHAB), the analytics (OpenSearch) and the more recent Artificial Intelligence (OpenWeb-UI, Ollama). Achieving this lofty goal requires the different systems to talk via MQTT. That's where this simple piece of Python code comes in.
Example of a motorized damper system directing air to the different zones.The air conditioner itself is a Daikin system (RZAS160C2V1) which is managed by the MyAir system.
In ceiling unit
The heat pump as installed at Electricbrain HQ is a 16kW unit
Outdoor unit
The resulting output into the MQTT server looks like this in MQTT Explorer:
advantage-air
aircons
ac1
info
aaAutoFanModeEnabled = True
activationCodeStatus = noCode
cbFWRevMajor = 10
cbFWRevMinor = 58
cbType = 1
climateControlModeEnabled = True
climateControlModeIsRunning = True
constant1 = 1
constant2 = 8
constant3 = 0
countDownToOff = 0
countDownToOn = 0
dbFWRevMajor = 17
dbFWRevMinor = 30
fan = low
filterCleanStatus = 0
freshAirStatus = none
mode = cool
myAutoModeCurrentSetMode = cool
myAutoModeEnabled = False
myAutoModeIsRunning = False
myZone = 2
name = AC
noOfConstants = 2
noOfZones = 8
quietNightModeIsRunning = False
rfFWRevMajor = 4
rfSysID = 0
setTemp = 24.0
state = on
uid = 0a6cc
unitType = 17
zones
z01
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 23.4
minDamper = 0
motion = 0
motionConfig = 1
name = TV Living
number = 1
rssi = 46
SensorUid = 0253a3
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 55
z02
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 23.5
minDamper = 0
motion = 0
motionConfig = 1
name = Kitchen
number = 2
rssi = 48
SensorUid = 0253a2
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 100
z03
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 22.5
minDamper = 0
motion = 0
motionConfig = 1
name = Lounge
number = 3
rssi = 45
SensorUid = 02539f
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 5
z04
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 22.0
minDamper = 0
motion = 0
motionConfig = 1
name = Bed Master
number = 4
rssi = 41
SensorUid = 0253a0
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 5
z05
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 23.0
minDamper = 0
motion = 0
motionConfig = 1
name = Office
number = 5
rssi = 32
SensorUid = 02539d
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 5
z06
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 21.8
minDamper = 0
motion = 0
motionConfig = 1
name = Bed Guest
number = 6
rssi = 40
SensorUid = 0253b1
setTemp = 23.0
state = open
tempSensorClash = False
type = 1
value = 5
z07
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 0.0
minDamper = 0
motion = 0
motionConfig = 1
name = Bath
number = 7
rssi = 0
setTemp = 24.0
state = open
tempSensorClash = False
type = 0
value = 100
z08
error = 0
firebaseSensorUpdateStore
followers = []
following = 0
maxDamper = 100
measuredTemp = 0.0
minDamper = 0
motion = 0
motionConfig = 1
name = Entry
number = 8
rssi = 0
setTemp = 24.0
state = open
tempSensorClash = False
type = 0
value = 100
myAddOns
myGarageRFControllers
myLights
myMonitors
myScenes
mySensors
myThings
myView
system
aaServiceRev = 14.142
backupId = f8a47145-4ea3-4ebe-b0af-709fa8d49c30
country = Australia
dealerPhoneNumber = 0123456789
deviceIds = ['NsPoVTT45bhpHhIoOdGCmBJ5JKp3', '1QDqoYi3V8RIq4OTmWhoqEZx0063', 'EKK9urSYQCVT6n5UdaKYISTpLEX3', 'thFkSKET2NM9oT5xgK4CTOMeOsl2']
deviceIdsV2
deviceNames
deviceNotificationVersion
drawLightsTab = False
drawThingsTab = False
garageDoorReminderWaitTime = 2
garageDoorSecurityPinEnabled = True
hasAircons = True
hasLights = False
hasLocks = False
hasSensors = False
hasThings = False
hasThingsBOG = False
hasThingsLight = False
isValidSuburbTemp = False
latitude = -40.7665875
lockDoorReminderWaitTime = 2
logoPIN = 4108
longitude = 147.29207689999998
membershipStatus = NotAMember
mid = 0a6cc
myAppRev = 15.1506
name = MyPlace
needsUpdate = False
noOfAircons = 1
noOfSnapshots = 0
postCode = 1234
remoteAccessPairingEnabled = True
rid = AcHdv5iYQtWxH2BvDgT8q2c0U332
showMeasuredTemp = True
splitTypeSystem = False
suburbTemp = 16.4
sysType = MyAir5
tspErrorCode = noError
tspIp = 192.168.1.29
tspModel = PIC8GS12
Communicating with MyAir
Talking to the Advange Air android pad's app, MyAir, is quite straight forward:
curl -X GET http://advantage-air:2025/getSystemData
where advantage-air is the local IP address on your LAN. The unit replies with lots and lots of JSON. That can be formatted by piping the returned text into jq:
curl -X GET http://advantage-air:2025/getSystemData | jq
Sending encoded JSON (officially known as percent-encoding) to the advantage-air controller can turn things on and off too:
curl -X GET http://advantage-air:2025/setAircon?json=%7B%22ac1%22%3A%7B%22info%22%3A%7B%22state%22%3A%22off%22%7D%7D%7D
AdvantageAir2MQTT Python program
The python program makes use of curl to talk to MyAir. The result is placed in a Python dictionary. From there it's converted to MQTT data commands. These are sent at a rate of about 10 per second. When everything has been sent the program disconnects from the MQTT server and goes to sleep for 60 seconds.
Most configuration parameters are gleaned from linux environment variables. This has the advantage of making things configurable from Docker.
It also need to be pointed out that the MQTT library version has to be less that 2.0.0 (i.e. version 1):
pip install "paho-mqtt<2.0.0"
Use this exact syntax to install the needed module (which at the time of writing in late 2025 is version 1.61) into your Python environment - Docker container details are described below.
#!/usr/bin/env python3
# advantageair2mqtt.py
#
# Queries AdvantageAir MyAir, convert and sends the results to MQTT server.
# Ctrl‑C stops the program.
import json
import os
import paho.mqtt.client as mqtt
import subprocess
import time
# Advantage Air tablet
advantageair_address = os.environ['ADVANTAGE-AIR']
# MQTT Broker details
broker_address = os.environ['MQTT_ADDRESS']
port = int(os.environ['MQTT_PORT'])
base_topic = os.environ['MQTT_BASE_TOPIC']
MQTT_USERNAME = os.environ['MQTT_USERID']
MQTT_PASSWORD = os.environ['MQTT_PASSWORD']
MQTT_ALLOWED_TYPES = [bool, bytearray, float, int, None, str]
# Define some empty variables
advantageair_data = {}
json_payload = ""
key_array = []
mqtt_messageid = 0
def on_connect(client, userdata, flags, rc):
"""Callback function for when the client connects to the broker."""
if rc == 0:
print("Connected successfully to MQTT Broker")
else:
print(f"Failed to connect, return code {rc}")
def on_publish(client, userdata, mid):
global mqtt_messageid
"""Callback function for when a message is published."""
mqtt_messageid = mid
def publish_recursive_dict(client, d, depth=0):
global key_array
topic = base_topic
if depth == 0:
key_array = [""]
else:
del key_array[depth:]
key_array.append("")
for key_element in key_array:
if key_element != "":
topic += "/" + key_element
for key, value in d.items():
print_topic = topic + "/" + key
if isinstance(value, dict):
key_array[depth] = key
publish_recursive_dict(client, value, depth + 1)
else:
if type(value) == list:
value = str(value)
if type(value) in MQTT_ALLOWED_TYPES:
client.publish(print_topic, payload=value, qos=1)
time.sleep(0.1)
print(f"{mqtt_messageid:5}: {depth:2}: {print_topic}: {value}")
else:
print(f"TypeError({type(value)}): {depth}: {print_topic}: {value}")
def get_advantage_air_data():
global advantageair_data
# Advantage Air details
result = subprocess.run(
[
"curl",
"--silent", # no progress meter
"-X", "GET",
advantageair_address
],
capture_output=True, # capture stdout+stderr
text=True, # decode bytes → str
check=False # we’ll handle non‑zero exit ourselves
)
advantageair_data = json.loads(result.stdout)
def main() -> None:
"""Infinite loop: print → sleep → print → …"""
get_advantage_air_data()
client = mqtt.Client("PythonJsonPublisher")
# Set username and password for authentication
client.username_pw_set(MQTT_USERNAME, MQTT_PASSWORD)
# Assign callback functions
client.on_connect = on_connect
client.on_publish = on_publish
while True:
get_advantage_air_data()
client.connect(broker_address, port, 60)
client.loop_start()
publish_recursive_dict(client, advantageair_data)
time.sleep(2) # pause for 2 seconds to allow all transactions to complete
client.disconnect()
print("Disconnected from MQTT Broker")
client.loop_stop()
time.sleep(60) # pause for 60 seconds
if __name__ == "__main__":
main()
Docker
Installing in to docker from the ElectricBrain website has a couple of difficulties. The main issue is the security and provenance of the container. How can anyone be sure they aren't installing code from some foreign government's intelligence agency. Well the answer is to use a container that has a traceable history. In this case the official Python container is used without modification. The various bits and pieces required to make things work are added/mapped in at container runtime by Docker. Nevertheless, a container has been uploaded to https://hub.docker.com from electricbrain. See the details below.
The only difficulty with this approach is installing the necessary PAHO-MQTT module. The path taken here was to map in a directory named /usr/local/lib/python3.14/site-packages-temp into a virgin running Python container (python:latest) which had its run command changed to /usr/bin/sleep 3600. After the container starts it immediately goes to sleep waiting for you to login to its commandline.
Upon logging in it's then a matter of installing the paho-mqtt module using the above command. This creates the needed /usr/local/lib/python3.14/site-packages directory complete with the module. The next step is to recursively copy the contents of the site-packages directory into the persistent directory on host's disk via the mapped in directory at /usr/local/lib/python3.14/site-packages-temp.
The Portainer template for the final docker container setup is shown below.
version: '3.7'
services:
AdvantageAir2mqtt:
image: my_registry:3999/python:latest
command: ["python", "-u", "/program/advantageair2mqtt.py"] # overrides default CMD
deploy:
mode: replicated
replicas: 1
placement:
constraints:
- "node.platform.arch==aarch64"
- "node.labels.public==true"
resources:
limits:
memory: 512M
reservations:
memory: 64M
environment:
- ADVANTAGE-AIR=http://advantage-air:2025/getSystemData
- MQTT_ADDRESS=myhost.localdomain
- MQTT_PORT=1883
- MQTT_USERID=AdvantageAirUser
- MQTT_PASSWORD=AdvantageAirPassword
- MQTT_BASE_TOPIC=advantage-air
restart: always
volumes:
- /myprivate/AdvantageAir2mqtt/etc:/etc
- /myprivate/AdvantageAir2mqtt/program:/program
- /myprivate/AdvantageAir2mqtt/site-packages:/usr/local/lib/python3.14/site-packages
Download the docker container from hub.docker.com
Head over to https://hub.docker.com and search for advantageair2mqtt.
>docker \
run \
-e ADVANTAGE-AIR=http://advantage-air:2025/getSystemData \
-e MQTT_ADDRESS=localhost.localdomain \
advantageair2mqtt:1.0.0-XXX64