A Home Assistant custom integration that finds the cheapest (or most expensive) upcoming hours based on
Nordpool electricity prices to save money automatically.

• Price-based scheduling — Automatically activates the cheapest/most expensive hours of the day
• Always-cheap threshold — Slots below/above a price threshold are always activated
• Always-expensive threshold — Safety cutoff to never activate above/below a certain price
• Price similarity threshold — Groups slots with nearly identical prices for more natural scheduling
• Rolling window constraint — Ensures minimum activity within any configurable time window (e.g., water heater must run at least 4 hours in any 24-hour window)
• Minimum consecutive hours — Prevents short on/off cycles by requiring a minimum run duration
• Excluded hours — Block a time range from ever being activated (e.g., to avoid grid fee peak hours)
• Multiple instances — Add one per appliance (water heater, floor heating, pool pump, etc.)
• Always on / Always off — Force all controlled entities ON or OFF via switches, bypassing the schedule
• Emergency mode — Keeps appliances running if price data is unavailable
• No helpers needed — All configuration is done through the integration's UI

• Home Assistant 2024.4.0 or newer
• Nord Pool integration, either the native addon or the HACS custom addon, installed and configured

1. Click the button above, or search for Power Saver in HACS
2. Click Install
3. Restart Home Assistant

1. Copy the custom_components/power_saver folder to your Home Assistant custom_components directory
2. Restart Home Assistant

Click the button above, or add it manually:

1. Go to Settings → Devices & Services → Add Integration
2. Search for Power Saver
3. Fill in the configuration:

4. Click Submit

To add another appliance, simply add a new service with different settings.

All scheduling parameters can be changed at any time via Settings → Devices & Services → Power Saver → Configure. Changes take effect immediately.

Each instance creates the following sensors:

Each instance includes two override switches:

• Always on — Forces all controlled entities ON regardless of the schedule. The status sensor shows forced_on.
• Always off — Forces all controlled entities OFF regardless of the schedule. The status sensor shows forced_off.

The two switches are mutually exclusive — enabling one automatically disables the other. Turn both OFF to resume normal scheduling. Switch states persist across Home Assistant restarts.

1. Price data — Reads prices from your Nord Pool sensor (today + tomorrow when available)
2. Excluded hours (optional) — Marks slots in the excluded time range as permanently off, removing them from scheduling
3. Slot selection — Selects the cheapest (or most expensive) slots from the remaining hours to meet your minimum active hours
4. Thresholds — Applies always-cheap (force on) and always-expensive (force off) price thresholds
5. Similarity grouping — Groups slots with nearly identical prices for more consistent scheduling
6. Rolling window (optional) — Ensures minimum activity within any rolling time window, activating additional slots as needed
7. Consecutive hours (optional) — Merges short active segments to prevent rapid on/off cycling
8. Updates — Recalculates every 15 minutes and immediately when new prices arrive

You can visualize the electricity price alongside the Power Saver schedule using the ApexCharts Card for Home Assistant. This gives you a clear overview of when the appliance is active and how it correlates with the price.

Replace sensor.heater_power_saver_schedule with your own schedule sensor entity ID.

type: custom:apexcharts-card
header:
  show: true
  title: Price 48t + Powersaver
now:
  show: true
  label: Now
graph_span: 2d
span:
  start: day
apex_config:
  stroke:
    width: 2
  dataLabels:
    enabled: true
  fill:
    type: gradient
    gradient:
      shadeIntensity: 1
      inverseColors: false
      opacityFrom: 0.45
      opacityTo: 0.05
      stops:
        - 10
        - 50
        - 75
        - 1000
  legend:
    show: false
  yaxis:
    - id: price
      show: true
      decimalsInFloat: 1
      floating: false
      forceNiceScale: true
      extend_to: end
    - id: usage
      show: true
      opposite: true
      decimalsInFloat: 0
      floating: false
      forceNiceScale: true
      extend_to: end
    - id: powersaver
      show: false
      decimalsInFloat: 0
      floating: false
      extend_to: now
series:
  - entity: sensor.heater_power_saver_schedule
    yaxis_id: price
    extend_to: now
    name: Price
    type: area
    curve: stepline
    color: tomato
    show:
      legend_value: false
    data_generator: |
      return entity.attributes.schedule.map((entry) => {
        return [new Date(entry.time).getTime(), entry.price];
      });
  - entity: sensor.heater_power_saver_schedule
    data_generator: |
      return entity.attributes.schedule.map((entry) => {
        return [new Date(entry.time).getTime(), entry.status === "active" ? 1 : 0];
      });
    yaxis_id: powersaver
    name: " "
    type: area
    color: rgb(0, 255, 0)
    opacity: 0.2
    stroke_width: 0
    curve: stepline
    group_by:
      func: min
    show:
      legend_value: false
      in_header: false
      name_in_header: false
      datalabels: false

The vast majority of this project was developed by an AI assistant. While I do have some basic experience with programming from a long time ago, I'm essentially the architect, guiding the AI, fixing its occasional goofs, and trying to keep it from becoming self-aware.