Ever had this happen?
A battery has been sitting at 1% for months and still works perfectly — but your dashboard says “CRITICAL”.
Or the opposite: a sensor slowly dies and you only notice once it goes completely offline.
That’s exactly the problem I’ve been working on recently — BatteryBrain.
The problem: Fixed percentage thresholds are not ideal. A device running on a coin cell at 1% behaves very differently from a Zigbee door sensor at 1%.
BatteryBrain solves this by automatically detecting all battery-powered devices in your Home Assistant setup, learning their discharge behavior, and classifying them into archetypes (linear, plateau-cliff, rechargeable, low-stable, etc.). Each type gets its own adaptive thresholds — reducing false alerts significantly.
Key features:
- Zero-config — install and it just works
- Imports existing Recorder history up to one year, so it’s immediately useful
- 4 summary sensors for dashboards and automations
- Detects when sensors stop reporting data
Available via HACS as a custom repository:
GitHub - larsbaum/battery_brain: Intelligent battery health analysis for Home Assistant · GitHub
This is still in beta — it runs quite stable in my setup, but I’m looking for testers and feedback. Feel free to open issues on GitHub.
Curious to hear what you think 