An untrained bystander streams their first-person view to a remote medic, who sees through their eyes and guides them — by voice, drawings, photos and a live map on the display — until help arrives.
In trauma, most preventable deaths happen in the first minutes, from causes a bystander could treat — catastrophic bleeding, a blocked airway. Trained medics can't always arrive in time. SaveVision lends their judgment over a wire.
From the phone or the glasses, the bystander reaches the operations centre. The glasses' point of view links automatically.
A licensed clinician watches the live POV, places the case on a map with the ambulance ETA, and triages.
The medic sends instructions, draws on the wound, sends reference photos, a live map — and approves AI proposals before they reach the wearer.
Mark exactly where to apply pressure or a tourniquet, on the actual limb.
Real first-aid images — tourniquet, CPR, recovery position — straight to the display.
A real map on the glasses, points, routes, and ambulance ETA — toggled on by the operator.
A vision model proposes procedure guidance; the clinician approves before it's sent.
Callers, situational map and task handling — synchronised in a single operator view.
Works on smart glasses with a display, or any phone camera as a fallback.
Video and guidance run over Matrix (Olm/Megolm) — not even the server can read them.
Self-hosted homeserver — casualty data stays on your own infrastructure.
One-way by design: the clinician decides, the bystander acts. Nothing auto-sends.
We're seeking pilot partners and sponsors. Try the demo, read the concept, and get in touch.