🔥 IN DEVELOPMENT

Hestia

A burnout early-warning system, dressed up as a fire dial.

Named for the Greek goddess of the hearth — watching the fire so it doesn't go out.

What it is

Instead of a clinical “you're burnt out” alert, Hestia uses a gentle five-point fire scale — Ash, Embers, Flame, Fire, Blazing — like checking how much fuel is left before it runs out. It weighs nine or more environmental dials (pollen, UV, moon phase, air quality, wind, travel distance, calendar density, and more) against your own history to give a heads-up before things get bad, not after.

How it works

  1. Onboarding walks through each dial with lag questions — how many days after a trigger does it usually hit you?
  2. Daily check-ins and phone-as-sensor data (like foreground-only sleep tracking) feed the model.
  3. A nearest-neighbour algorithm compares today's fingerprint against your own past patterns across several window lengths — not a generic average across all users.
  4. Near-misses are marked as negative examples and recent data is weighted more heavily, so the model keeps adapting to you specifically.

Why it exists

The problem

Most burnout tracking is either too clinical to want to open, or it only tells you after the crash has already happened. Hestia is trying to catch it earlier, and say it kindlier — the fire scale is a metaphor everyone already understands.

Under the hood

All data — including medication adherence — is encrypted at rest on-device.

React Native / Expo SQLite (per-row tables) AES-256-GCM via expo-crypto expo-secure-store Nearest-neighbour prediction model
Burnout flag (ember) · #D4612A
← Back to the whole suite