Better Thermostat

For more info visit: https://better-thermostat.org/

Requirements

• Minimum required Home Assistant version: 2026.9.0 (continuously tested against Home Assistant )
• Minimum required Python version: 3.14.2 (inherited from Home Assistant; the integration uses syntax that older interpreters reject)

Companion UI

We've created a companion UI element which can display more information than the default thermostat element in Home Assistant. Check it out via HACS: better-thermostat-ui-card

• If you have a question or need help please create a new discussion or check if your question is already answered
• If you have a suggestion, found a bug, or want to add a new device or function create a new issue
• If you want to contribute to this project create a new pull request

Features

This integration brings some smartness to your connected radiator thermostats setup:

• Uses a temperature sensor far away from the radiators to measure the real room temperature
• Makes your TRVs fully compatible with Google Home
• Let your windows disable your heating (avoid programming this via an automation)
• Your weather forecast provider will turn your heat on/off
• Or an outside air temperature sensor can do this as well
• Does some valve maintenance automatically, to avoid them getting stuck closed over the summer
• Group multiple TRVs to one (e.g. for a room with multiple radiators)
• Enhance the default TRV Algorithm with some smartness to reduce energy consumption
• Dynamic preset temperature learning & persistence (baseline/"no preset" remembers the last temperature you set and survives restarts)
• Advanced Control Algorithms: Choose between AI Time Based, MPC, MPC v2, PID, TPI, Aggressive, External Sensor Offset Only or No Calibration.
• Selectable Presets: Configure which preset modes are available for your thermostat during setup.

Advanced Control Algorithms

Better Thermostat now supports multiple advanced control strategies to optimize your heating:

• MPC (Model Predictive Control): Uses a physical model of your room and radiator to predict future temperature changes and optimize valve opening.
• PID Controller: A classic Proportional-Integral-Derivative controller that learns your room's characteristics to maintain a stable temperature. It features auto-tuning (currently in beta) to automatically find the best parameters (Kp, Ki, Kd) for your room.
• TPI (Time Proportional Integral): Turns the distance from the target into a duty cycle and holds the valve at that opening, which keeps the temperature stable and reduces overshoot.
• AI Time Based: Uses a custom algorithm based on simple measurements and calculations (not actual AI) to calculate the required heating power and adjusts the TRV calibration to achieve it. This improves upon the standard TRV internal algorithm. It is the default.
• MPC v2 (experimental): A predictive controller with a Kalman observer and a quadratic-program optimiser. It writes the valve directly with Direct Valve Based and otherwise works through the target temperature or the offset.
• Aggressive: Pushes the TRV harder while heating for a faster warm-up, at the price of more overshoot.
• External Sensor Offset Only: Corrects the TRV's internal sensor with your room sensor and nothing more.
• No Calibration: Passes your target temperature to the TRV unchanged.

Not sure which to pick? A quick guide:

Algorithm	Best For	Trade-off
TPI	A robust default — steady across room types, gentle on valve & battery, good with multiple radiators	Slightly looser setpoint tracking than PID/MPC
PID	Tight setpoint tracking in a stable, well-characterised room	More valve activity; varies more by room; autotune is still in beta
MPC	Anticipating solar / known disturbances when the room is modelled well	Most valve activity; can over-react with several radiators; still in beta
AI Time Based	Quick setup, no tuning	Looser comfort, weaker in complex rooms

Start with AI Time Based (the default) or TPI. Try PID for tighter tracking in a stable room, or MPC for disturbance anticipation. The algorithm documentation describes every mode.

These modes can be selected in the advanced configuration of the device.

Preset Temperature Configuration

Preset temperatures are now fully configurable via dedicated number entities.

How it works:

1. During setup or configuration, you can select which Presets you want to enable for this thermostat.
2. For each enabled preset mode (e.g. Eco, Comfort, Sleep), a corresponding number entity is created (e.g., number.better_thermostat_preset_eco).
3. These entities are located in the Configuration category of the device.
4. You can adjust the temperature for each preset directly using these number sliders.
5. The values are automatically persisted across Home Assistant restarts.
6. Changing a preset temperature via the number entity immediately updates the thermostat if that preset is currently active.

Default starting values:

Away:            16.0 °C
Boost:           24.0 °C
Comfort:         21.0 °C
Eco:             19.0 °C
Home:            20.0 °C
Sleep:           18.0 °C
Activity:        22.0 °C

Which hardware do we support?

We support all thermostats which are compatible with Home Assistant as long as they are shown up as a climate entity.

Integrations that are tested

• Zigbee2Mqtt
• Deconz
• Tado
• generic_thermostat

How to setup

Install this integration via HACS or copy the files from the latest release

Configuration details can be found in the documentation or on our website: better-thermostat.org

Some nice-to-know config tips for the configuration.yaml

Example Window/Door - Sensor config

group:
  livingroom_windows:
    name: Livingroom Windows
    icon: mdi:window-open-variant
    all: false
    entities:
      - binary_sensor.openclose_1
      - binary_sensor.openclose_2
      - binary_sensor.openclose_3

Combine multiple TRV to one (Group)

No worry, Better Thermostat supports grouping out of the box

---

Contributing?

checkout the CONTRIBUTING.md file

☕ Support

If you want to support this project, you can ☕ buy a coffee here.