Moodbound

Seven-point mood tracking with an on-device safety engine. Apple Health integrated. Open source, no cloud, no telemetry. Free, no subscriptions.

Moodbound tracks your mood on a seven-point scale from depression to mania. It reads sleep, heart rate, and activity from Apple Health, looks for patterns, and tells you when your trend is shifting.

Built for bipolar disorder. Useful for anyone who wants to take their mood seriously.

WHAT IT DOES

A check-in takes under a minute: mood, energy, sleep, irritability, anxiety, with optional medications and triggers. Instead of asking the same things every day, Moodbound looks at your history and picks what's worth asking next.

Sleep hours pull from your watch automatically. Resting heart rate, HRV, step count, and mindful minutes ride along, adding context you'd never track by hand. Each entry gets written back to Apple Health as a State of Mind sample.

PATTERN DETECTION

A Bayesian safety engine runs against your history every time you open the app. It combines mood trajectory, sleep, HRV, activity, and medication adherence into a single risk assessment at four levels: none, elevated, high, critical.

Change-point detection catches the moment your trend bends. Conformal prediction gives you a forecast with honest uncertainty intervals, so you can see where the model is less sure. Wasserstein drift scoring measures how far your recent days have wandered from your baseline.

When the numbers move, Moodbound tells you what changed, what evidence supports the score, and what to do next.

SAFETY PLANNING

Your safety plan lives in the app: warning signs, coping strategies, the steps to take when you're not OK. Support contacts are one tap away. The safety engine surfaces your plan automatically when risk climbs.

PRIVACY

No account. No cloud sync. No analytics. No telemetry. All data stays on your device in encrypted storage. Export the full record as JSON whenever you want. The source code is public on GitHub.

Free, with no subscriptions, no in-app purchases, no paywall. If you'd rather build it yourself, the source is on GitHub.