Inspect the room. Not the person.

Inspect the room. Not the person.

Shared housing often breaks down because messes go unattributed, leaks go undetected, damage goes unresolved, and operators have no fair way to hold anyone accountable. The obvious answer is also the wrong one: putting cameras everywhere destroys trust, creates legal exposure, and turns housing into surveillance.

That is the line. We measure access logs, room status, sensor data, and after-vacancy condition. We do not record people in private spaces. Every consequential decision routes to a human, with an appeal on record.

Modular by design. Start small. Scale into it.

Shelterfy hardware is delivered as five interlocking kits. Operators do not need everything on day one. The minimum viable deployment is the Basic Accountability Kit. Robotics arrive only after the operating model is proven.

Financed as property infrastructure

Shelterfy hardware is not a gadget expense. It is risk-reduction and operations infrastructure that funds itself through property acquisition budgets, renovation CapEx, sponsor contributions, insurance incentives, operator technology fees, and the maintenance savings the system itself produces. When the math works, the property pays for the system. When it doesn't, the deployment doesn't ship.

Four surfaces. One source of truth.

Every Shelterfy user gets a tailored surface. All four sit on the same operational data layer, so a sensor event in the field shows up in the operator dashboard, the resident's app, the sponsor's impact report, and the insurance risk panel without a separate integration project per stakeholder.

Operator dashboard

Real-time control of the property. Occupancy overview, vacant-room scan queue, room status map, property health score, device status, maintenance alerts, incident review queue, restorative actions, sponsor reporting, audit log. Privacy-first labels everywhere so a property manager can answer hard questions in the open.

Resident app

Built to feel empowering, not surveillant. Digital ID, check-in / check-out, access QR / NFC, current booking, trust score, access level, active tasks, restorative action plan, premium amenities, job opportunities, ownership progress, savings progress, appeals, and a clear "what data is collected" section.

Sponsor portal

Campaigns, donations, product contributions, property sponsorship, resident-credit sponsorship, impact reports, hiring pipeline, geo-targeting, brand placement review. Sponsor metrics roll up by nights of housing supported, meals or hygiene kits provided, jobs created, ownership credits matched, and property-health improvement. Outcome reports without exposing personally identifiable resident data.

Insurance + risk portal

Property-level risk reports for insurance partners. Property health score, maintenance response time, leak history, fire / safety alerts, damage frequency, inspection coverage, claims prevention. The portal is engineered to expose risk-relevant aggregates, not resident behavior.

Two rails. One human in the loop.

Shelterfy uses a two-rail AI model: a Physics Rail for hard rules that cannot be violated, and a Machine Learning Rail for classification, prediction, and recommendations. Neither rail decides serious consequences for a resident. Humans do.

Physics Rail (hard rules)

• A private-space scan cannot start unless vacancy is confirmed.
• A camera cannot be assigned to a bathroom, bedroom, shower, or private sleeping space.
• A scan in progress aborts the moment a presence signal returns.
• A resident cannot be suspended, charged, or removed without human review.
• A sponsor report cannot include resident PII without explicit consent and policy match.
• A maintenance alert must fire if a moisture, smoke, or air-quality threshold is crossed.

ML Rail (recommendations, never final say)

• Classify room condition after a scan.
• Detect spills, broken items, debris, water presence, and unusual conditions.
• Predict maintenance risk and prioritize cleaning workflow.
• Recommend a restorative action that fits the incident.
• Match residents to available marketplace work and training opportunities.
• Flag residents who may need support before an incident escalates.

Bronze, Silver, Gold. Immutable trail, single source of truth, disposable reporting.

Shelterfy uses a medallion data architecture. Every device event lands in Bronze first, immutable. Silver is the normalized operational layer that powers the dashboards and the resident app. Gold is reporting and analytics, rebuildable from Silver at any time.

The full sequence, step by step.

This is the operating loop a Shelterfy property runs every time a private space is used. Nothing in the loop violates the core principle. Every step has a privacy guarantee.

1. Resident enters a private room. Smart lock, QR, NFC, or door sensor logs access.
2. No camera or scan is active during use. The room belongs to the resident.
3. Resident exits. Door sensor detects exit.
4. Occupancy + presence sensors confirm the room is empty.
5. A vacancy-confirmation hold runs. The system double-checks no presence is detected.
6. Telescoping arm or ground rover performs the post-use inspection. Floor, fixtures, surfaces, environmental readings.
7. Edge AI classifies the room condition.
8. Minor findings post a cleaning task. Serious findings route to the human review queue.
9. Confirmed issues link to access logs, timestamps, and room status, so attribution is fair.
10. Resident is notified with a restorative action path. Cleaning shift, repair, training, mediation, partial repair contribution, or appeal.
11. Operator dashboard updates property health, maintenance records, and the audit trail.

Built to be defended. Not retrofitted to apologize.

Every objection Shelterfy will face is anticipated in the system itself. The defense is structural, not rhetorical.

Six phases. Software first. Robotics last.

Shelterfy is engineered to prove the operating model in software before it spends a dollar on robotics. Hardware complexity scales in only after the workflow is real.