Cloud Status

Realtime sync architecture for devices, events, settings and training.

Realtime backend not connected yet
Realtime sync
Mocked
listeners ready
Write latency
150 ms
Events stored
1,284
Auth
Email + Google
planned providers

Database structure

/crysense
  /devices
    /device_001
      status, battery, wifi, microphone,
      model_version, last_seen
  /events
    /event_id
      timestamp, prediction, confidence,
      duration, mode
  /settings
    sensitivity, confidence_threshold, cooldown
  /training
    datasets, models, experiments

Access rules

  • Each account can read and write only its own devices, events and recordings.
  • Device writes are authenticated with a per-device token, never a shared key.
  • Training datasets keep consent and license metadata alongside every sample.
  • Deleting an account removes all cloud recordings and event history.

Realtime event payload

{
  "prediction": "Hungry-like",
  "confidence": 0.82,
  "timestamp": "2026-09-08T20:42:11Z",
  "latency": 420
}

When this arrives, the dashboard status, cry history, analytics and notification center all update together with smooth transitions.

AI predictions are estimates and should always be interpreted together with context and normal caregiving judgment. CrySense AI is not a medical device.