One simulated minute every 1.5 seconds · every event goes through the real API
Stops it reporting and puts a new pairing code on its face, ready to type into a different account.
Waiting for the first evaluation…
No alerts. That is the correct answer most of the time.
A whole story, played against this household in simulated time. Every event goes through the real ingestion, the real rules and the real escalation policy — and you can take over at any point. Omblit does not diagnose: each story is played through the signals a watch and a few sensors can genuinely observe, and the alert you see is the one the pipeline produced.
Loading scenarios…
The signals that reach a care team fastest. A hard fall with no movement afterwards and a deliberate SOS both escalate without waiting; a possible fall asks the wearer first.
The heart-rate rule compares against this wearer's own baseline, not a fixed threshold, so the same reading is a finding for one person and unremarkable for another.
The cases that separate a working pipeline from a plausible one. Each of these must be refused or collapsed rather than believed.
A missing signal is not an emergency. Pull devices off the network and the score is marked incomplete rather than reported as calm.
The safety check is answered on the watch, not here — the face takes over when one is waiting. This is the same button.
Every account the development API will hand out a session for. Picking one takes over its devices, so the caregiver app shows the simulated watch as that person's live data. Their profile is left exactly as it is — the baselines the engine reads stay theirs.
Loading households…
Nothing attached yet.
Puts a code on the watch, as real hardware does. Type it into the caregiver app, or use “Add it for me” on the watch face.
Stillness in these hours is not a finding. It is what keeps a sleeping house from looking like an emergency — and it is scoped to the inactivity rule, so a fall at 3am still escalates.