Care OS robot assisting a nurse in a hospital corridor
Care OS On-device
The trust layer for healthcare robotics

Robots your staff
actually trust.

The behavioral intelligence layer that makes any healthcare robot move safely and politely around people.

The bottleneck

Healthcare is ready. Robots are ready.
Humans are not.

Social acceptance is the bottleneck, not navigation.

01 Staff shortage

Nurses lose hours to errands a robot could run.

02 Robots behave wrong

They block halls and startle patients. They don’t belong yet.

03 Trust collapses

One bad encounter and staff stop using it for good.

“The robot works, but people won’t use it. Trust is gone.”
Facility Manager, Swiss hospital
The solution

Care OS

A behavioral intelligence layer for robots. Hardware-agnostic · Human-aware · Healthcare-ready.

01 Socially-aware behavior

Slows, yields and gives way like a considerate colleague.

02 Context-adaptive interaction

Reads the room and adapts how it moves and signals.

03 Care-compliant operation

On-device, respects clinical norms, integrates on your terms.

Nav2 gets the robot through the corridor.
Care OS gets it accepted.

The pipeline

How Care OS works

Raw sensors in, socially-aware motion out. In milliseconds.

01 Sense

Existing sensors, read as anonymous geometry.

02 Understand

A live read of people, space and intent.

03Behavior Engine Decide

Picks a behavior, with an explicit reason.

04 Act

Caps speed, pauses, yields or re-routes.

Built for hospital constraints
01 On-device

Split-second decisions on the robot. No cloud.

02 No cloud required

No patient data leaves the building.

03 Optional integration

Plugs in only where you allow.

04 Plug and test

Evaluate on a real shift, in a day.

We don’t build robots. We upgrade their intelligence.

Talk to us
World Model Recorded mission
Privacy by architecture

Understands the room.
Keeps it private.

It anticipates how a corridor will unfold, the way an experienced nurse reads a busy ward.

Anonymized geometry, never images. Video never leaves the building.

How it thinks
Today: a deterministic rulebook

Reacts to what’s in front of it, one moment at a time. Safe, but stiff, and blind to what’s next.

Where it’s heading: a learned world model

Predicts how the scene will move and plans seconds ahead. Sharper with every mission on your floor.

Always shielded by the proven safety controller. It proposes; the rulebook keeps the brakes.

From reacting to what is, to anticipating what’s next.

Safety stays in charge. The learning never gets the brakes.

In the corridor

Where Care OS delivers value

01 Assist nurses

Errands off the ward, so staff focus on patients.

02 Internal logistics

Samples, supplies and meals, moved between departments.

03 Staff interaction support

Greets, guides, signals intent.

04 Interactive engagement

A light, friendly presence on the ward.

Designed with Swiss hospital staff, first validations begin 2026.

See it on your ward

A 1-day on-site evaluation

Free of charge. No integration required. Request your pilot
01 Setup

30–60 min to bring a robot online on your floor.

02 Test scenarios

2–4 h running real corridor scenarios with your team.

03 Live observation

Staff-in-the-loop: watch how people actually react.

04 Wrap-up

30-min review with KPIs, and optional follow-up.

Who we are

Research × Innovation × Business Execution

Salim Benamira
Salim Benamira Founder & CEO
Anatole Conrad
Anatole Conrad Robotics Engineer & CTO
Camille Lavilla
Camille Lavilla Robotics Engineer
Prof. Dr. Yulia Sandamirskaya
Prof. Dr. Yulia Sandamirskaya Scientific Advisor
Ecosystem

Built within a strong ecosystem

  • NVIDIA Inception
  • Microsoft for Startups
  • Stereolabs
  • Innosuisse
  • Tenity
  • ZHAW RobotCare
  • BFH
The ask

Join the future of healthcare robotics