The AI decides, and then acts
KinetiCura evaluates every deviation against the person's own baseline and notifies your prioritized circle of care on its own, with an alert SMS notification, without a human analyst in the loop.
Why KinetiCura?
Because a system that does not know you, cannot care for you
KinetiCura was built around a single idea: safety technology should understand the person it is watching over, and it should act without waiting for someone to interpret the data first.
The difference
What Makes KinetiCura Different
Plenty of systems can raise an alarm. Other AI monitoring products put a support team behind the software: analysts watch the data, spot the pattern, and pass a message to the care team. KinetiCura closes that loop itself, because the intelligence sits in the system, not in a control room waiting to interpret what the data means.
Trends are read continuously, not reviewed weekly
Gradual changes — slower gait, longer periods seated — are surfaced, so care can be adjusted while the change is still small.
Identity-aware in a shared home
Person re-identification keeps each individual's baseline correctly attributed as they move between rooms and cameras, and household members can be distinguished from visitors — so a busy family home does not degrade into noise.
A baseline per person
Thresholds are personal, not one-size-fits-all. What is alarming for one resident is an ordinary Tuesday for another, and KinetiCura knows the difference because it learned it.
Runs at the edge
Inference happens locally on affordable edge hardware beside the cameras. Only events, never footage, reach the cloud.
Privacy is structural, not a setting
Only skeletal pose markers are taken from each frame before it is discarded. There is no stored footage to leak, subpoena or misuse, and inference runs locally rather than in someone else's data centre.
Autonomous facial registration — but no identification
Faces are automatically detected and registered as anonymous individuals. This is what lets KinetiCura tell one household member from another and filter out visitors — it does not build a named identity record of who anyone is.
Guaranteed alert notification
An alert event is sent to the cloud, where the sending of the alert itself is managed and monitored. Delivery to your circle of care does not depend on a single app, device or connection staying up.
GDPR/POPIA compliance and data privacy by design
Data minimisation is designed in from the start, not added afterwards. The most sensitive data — recognisable video — is never created, and consent records, lawful basis and audit trails are built to meet both UK/EU GDPR and South Africa's POPIA.
Technology
Edge intelligence, cloud oversight
Detection happens beside the camera. Only events — never footage — reach the cloud, where families and care teams get their dashboard, alerts and reporting.
On-site edge layer
Runs locally, in real time
- Pose estimation extracts skeletal keypoints — joints and torso angle — to model posture and movement.
- Multi-object tracking holds a persistent identity for each person across frames, through brief occlusion.
- Re-identification re-associates a person after they leave and re-enter view or move between cameras.
- Facial recognition registers household members so monitoring applies to the right people, with visitors monitored for acute emergencies while on the premises.
- Event state machine applies debounce logic and clinically-informed thresholds so sitting, bending or lying down by choice are not mistaken for a fall.
Cloud layer
Everything the humans need
- Carer dashboard — web and mobile views of live status, activity history and alert logs for each monitored person.
- Alerting services — alert SMS notification to a nominated and prioritized circle of care.
- Encrypted records — household, baseline and event history used to personalise thresholds and support longitudinal reporting.
- Insight and reporting — anomaly trends aggregated over time to surface patterns such as gradually declining physical mobility and/or environmental risks.
- System monitoring — uptime and camera connectivity tracked across every deployment, so a silent camera is never mistaken for a quiet night.
See it read a real room
See a demonstration of baseline learning, fall detection and automatic notification of your circle of care. Questions about privacy and false alarms? Ask us directly.