Steward Health
Health is the proving ground. Continuity is the product.
Health is the first domain because useful information is already fragmented across watches, medical systems, fitness platforms, sleep data and personal observation. Steward is not ultimately a health application — health is where the architecture proves itself.
Steward is currently in development. The initial prototype is focused on proving persistent continuity across health and adjacent context.
The fragmentation
The pieces already exist.
Almost nothing here is missing. Each source is real, each one is partly useful, and none of them can see the others.
The missing element is not more data. It is continuity between what already exists.
Apple Health
HealthKit records on the phone
Health Connect
Android's aggregation layer
Wearables
Strain, heart rate, readiness
Sleep
Nightly duration and quality
Activity
Training and daily movement
Weight
Trend over months, not one reading
Medical records
Structured clinical history
Calendar
What the weeks actually looked like
Goals
What the person is working toward
Observations
What the person notices themselves
The pieces exist. The continuity between them does not.
A real question
“Why has my energy been lower lately?”
No single application can answer this. Steward assembles the sleep decline, the evening schedule change, the shift in activity, the recovery signals and the personal context — then shows its work.
You
Why has my energy been lower lately?
Assembling context
- Sleep ↓ 6h12m avglast 3 weeks
- Evening commitments ↑ 4 → 9 / weekcalendar
- Recovery ↓ trendingwearable
- Activity ↑ volumetraining log
- Schedule changedsince 28 Jul
Steward
Your sleep has declined over the last three weeks, but that change appears to have started after your evening schedule became significantly busier. Your exercise volume also increased during the same period while recovery declined.
Those changes happened together, but they do not establish a cause.
Would you like me to compare this period with the last time your energy dropped and help identify what was different?
Key factors
- — Sleep ↓ 6h12m avg
- — Evening commitments ↑ 4 → 9 / week
- — Recovery ↓ trending
- — Activity ↑ volume
- Sleep duration · wearable · daily, last 90 days
- Evening events · calendar · permissioned, read-only
- Training volume · fitness platform · weekly totals
- Prior energy dip · Steward memory · Nov 2025
- Provenance retained · permissioned sources only
Illustrative Steward concept — product currently in development
Why do you think that?
Answerable with sources attached
What changed?
Answerable with sources attached
What happened before?
Answerable with sources attached
What Steward does not do
Steward does not diagnose.
Its role is narrower, and more durable:
- Organize context across sources and time
- Surface changes worth noticing
- Preserve longitudinal history
- Help prepare better questions
- Support better conversations with qualified professionals
Interpretation belongs to the person and to qualified professionals. Steward makes that conversation better informed.
Crossing the first boundary
Schedule explains what health data cannot.
The first connection outside health is the calendar. A busier evening schedule explains a sleep change that no health application can see on its own.
Health sees
- Sleep
- Activity
- Weight
- Records
Sleep duration down 46 minutes across three weeks.
Calendar sees
- Meetings
- Travel
- Schedule density
- Routine changes
Four late commitments per week, starting the same period.
Steward sees
- Sleep ↓
- Evening load ↑
- Recovery ↓
- Prior dip · Nov 2025
One pattern, with both sources attached and nothing overwritten.
Next
Trust is part of the architecture, not a policy statement.