Hi everyone,
I would like to introduce my new Home Assistant project: ThermoSmart.
ThermoSmart is a fully local, self-learning heating controller that continuously learns how rooms heat up and cool down over time. It combines this knowledge with weather conditions, forecasts and occupancy information to make increasingly informed heating decisions.
The goal is not only to control thermostats, but also to understand the thermal behaviour of a building and adapt heating strategies accordingly.
Key Features
- Observation Mode
- Outcome Scoring
- Adaptive Preheating
- Heat Rate & Cool Rate Learning
- Weather & Forecast Integration
- Multi-Zone Support
- External Temperature Sensors
- Direct Valve Control (where supported)
- Setpoint Boost Fallback
- Fully Local Operation
Observation Mode
One of the core features of ThermoSmart is Observation Mode.
ThermoSmart can learn how your home behaves before taking over active heating control. This allows users to evaluate the system, collect learning data and understand how ThermoSmart would behave before handing over control.
Current Status
Current release:
ThermoSmart v1.0.1-beta.5
The integration is now feature-complete and I am looking for beta testers and feedback from different Home Assistant environments.
Looking For Testers
I am especially interested in testing with:
- Zigbee2MQTT
- ZHA
- Z-Wave
- Matter
- Sonoff TRVs
- Danfoss Ally
- Tuya Thermostats
- Other Home Assistant climate entities
GitHub
ThermoSmart Integration
ThermoSmart Dashboard Card
Feedback, ideas and bug reports are very welcome.
Thanks for taking a look!