Lives in the room.
A quiet local AI box that sits on a shelf at home. Nothing runs in a data centre, and nobody can open it like a camera.
The dangerous fall is the one nobody sees.
An unwitnessed fall can leave someone on the floor with no way to call for help. A wearable only works if it is worn and can be activated. A normal indoor camera only helps if someone happens to be watching it.
ilonai closes that gap without putting anyone behind a screen: the detection happens automatically, and the only thing a person receives is a message when help may be needed.
In Switzerland, the BFU reports (September 2026) that more than 1,700 people die from falls each year, 95% of them aged 65+, with nearly 6,500 people seriously injured and around CHF 1.8 billion in economic costs.
On a normal day
- AI processes locally
- A short rolling buffer is overwritten
- No image leaves
- No one can watch live
- No activity history is created
- Only a tiny system-health heartbeat leaves
If nothing happens, no image leaves the system.
If a possible fall happens
- AI flags an emergency locally
- One confirmation image can go to a designated guardian or caregiver
- A short pre/post incident clip can stay on the local box
- That person decides what to do
Someone is contacted only when help may be needed.
We deliberately do less.
ilonai does not need to know how you spend your day to detect a possible fall.
Other care technology measures sleep and movement, and builds a picture of someone’s routines. ilonai has a narrower job: detect a possible fall, get help, and otherwise leave the person’s life alone.
No live feed
No audio
No face recognition
No sleep or activity reports
No cloud video
No ilonai staff viewing
Powerful AI. Kept local.
Typical network cameras have to run very small models inside the camera, or send imagery to a cloud service. ilonai uses a dedicated local AI computer, so the detector can work with richer context while camera images stay inside the home or building.
More compute, in other words, aimed at one question instead of a wider view of your life.
Its own private safety network.
Cameras connect to the local AI box over a closed network of ilonai’s own, wired wherever possible. Where a cable can’t be run, a camera joins that same closed network wirelessly. Either way the cameras never touch your household Wi-Fi. A separate cellular connection carries system status and emergency alerts, and nothing else.
No camera cloud account. No dependence on household Wi-Fi. No port forwarding.
Detection keeps running locally even when the external connection drops.
A safety system should tell you when it stops working.
Silence should mean “everything is fine,” not “the system stopped working.”
ilonai watches its own health: whether the cameras are reachable and sending usable images, whether the AI computer and the cellular heartbeat are alive, and whether the power has been out for a while. If protection is interrupted, guardians or staff hear about it.
Why ilonai is different
| Wearable button | Normal smart camera | In-camera fall AI | ilonai | |
|---|---|---|---|---|
| No action required after a fall | No | Depends | Yes | Yes |
| No wearable to charge or remember | No | Yes | Yes | Yes |
| No live human surveillance | Yes | Usually not the design | Varies | Yes, by design |
| No normal cloud video | Yes | Varies | Varies | Yes |
| Dedicated local AI compute | No | Usually cloud/vendor | Limited | Yes |
| Does not build a daily behavior profile | Yes | Varies | Varies | Yes, by design |
“Varies” means products in that category differ. We do not claim every competing product works the same way.
Choose your setting
For someone living at home
Independence without constant checking. Family cannot open a live camera feed. They hear from the system only when it believes help may be needed.
For homeFor home-care services
You're there for an hour. ilonai covers the other 23. Families contract with us; your organisation can join as a guardian if it wants to.
For home-care servicesFor care homes
Scale fall detection without scaling surveillance. Cameras on a closed network, shared local AI, event-only visual access.
For care homesThe home pilot is underway. Evidence comes next.
The first pilot started on 5 October 2026 in Ilona’s home, the reason ilonai exists. There is room for four more, then we go commercial.
The detector works in our own testing, on recorded test material and on staged falls in the founder’s home. From the pilot we will publish what a few homes can actually produce: false alerts, uptime, and staged detection in situ. Real falls are rare, so they will be reported as cases when they happen, not as a rate.
We will not present development targets as real-world performance results.
Explore
How it worksA normal day vs. when help may be needed
TechnologyLocal AI, private camera network, alerts
Privacy & TrustWhat leaves the home, and what never does
Privacy explainerA short page for residents and families
EvidenceWhat we can say now, and what comes after the pilot
Pilot programFirst pilot started 5 October 2026; room for four more
For homeIndependence without constant checking
For home-care servicesSpitex and private home care
For care homesDetection without a live camera wall
AboutIlona, Zug, and why the product stays narrow
FAQShort answers to common questions
For investorsMarket, model, and where we are
ContactWrite to info@ilonai.ch
