Trust & safety
You are letting a stranger into your home on your worst morning.
Vetting
Ten gates. All of them, every time.
An applicant cannot be activated until every row below passes. This is enforced by a single function in our codebase that returns the list of unmet gates — not by a checklist someone works through.
| Gate | What it requires | How it's verified |
|---|---|---|
| Identity | Government photo ID plus a selfie liveness check. | Automated, before anything else |
| Criminal background | Multi-jurisdiction criminal search, national sex offender registry, SSN trace, and county-level records for seven years. | Through Checkr |
| Driving record | Motor vehicle record check. | Through Checkr |
| References | Two professional references, contacted by a person and logged. | Manual — we actually call |
| Credentials | CPR and paediatric first aid for child care. CPR plus CNA or HHA for adult care where the state requires it. | Verified at source |
| Health screening | TB testing where the state requires it. Immunisation records for centre placements. | Document verification |
| Interview | A structured video interview against a scored rubric, with a real person. | Recorded and logged |
| Training | Scope of care, mandatory reporting, emergency procedures, infection control, plus dementia basics for the adult track and safe sleep for the infant track. | Re-certified annually |
| Work authorisation | I-9 or W-9 depending on the engagement model in that market. | Before activation |
| Payout setup | Verified bank details through Stripe Connect. | Before first job |
Staying vetted
Vetting isn't a one-time event.
Most of the risk in this category isn't the caregiver who was never checked — it's the one who was checked two years ago.
Background check re-run. Expiry automatically pauses the caregiver — no exceptions, no grace period.
Criminal monitoring where available, so a new charge surfaces between annual checks.
Credential expiry warnings, then automatic pause on the expiry date.
Any rating below 3, any high-severity incident, or two no-shows in 90 days triggers automatic review and a pause pending decision.
Scope of care
Non-medical only. This is a hard boundary.
It's written into caregiver terms, family terms, the intake questions, and the software itself. The adult intake screens for every item below and tells your employee plainly when we can't help, with referral options — rather than accepting a booking we can't safely staff.
- No medication administration, dispensing or injections — reminders only
- No wound care, catheter care or ostomy care
- No transfers requiring a lift device
- No transporting a care recipient in any vehicle
- No children from different families with one caregiver in a home
- No overnight in-home child care at launch
When something goes wrong
What actually happens, in order.
- 1
Report
An always-visible button in both the caregiver and family apps. Severity is defined in plain language so people report rather than hesitate.
- 2
Page
High severity or any allegation of harm pages our on-call immediately. Not a ticket queue — a phone ringing.
- 3
Contact
A coordinator reaches the family within fifteen minutes.
- 4
Pause
The caregiver is automatically suspended pending review for any allegation of harm. Before an investigation, not after.
- 5
Investigate and report
Documented investigation, resolution, and mandatory external reporting where the law requires it. Records are append-only.
Questions
Including the uncomfortable ones.
Can anyone override the vetting process?
How current is a background check?
What can a caregiver never do?
What happens if there's an incident?
Are caregivers mandatory reporters?
What insurance covers a booking?
Can a family request a caregiver by race or religion?
Ask us the hard version of any of this.
We'd rather answer it now than during your security review.