Operator Monitoring
The task is already monitored. EmoPulse adds the person carrying it out.
About EmoPulse
EmoPulse builds the human-state layer that existing platforms are missing.
Platforms keep getting better at reading the world around a person. What is missing is the person. We think the next step is for them to follow the person too: how their state changes over time, measured against their own normal. We do not build a platform of our own. We build that one layer, so that it can be added to what already exists.
A patented software layer, on the device by default.
Part I: The Layer
A system that sees still misses the person
A system that sees the road, the machine or the room still has no idea who is in it or how they are doing.
The missing piece is not another sensor. It is a way of reading the person from what the system already has, and of following how fatigue, workload, attention and behaviour change in that person over time.
EmoPulse is not an emotion classifier and not a standalone product.
EmoPulse adds that missing human-state layer. It derives human-state signals from ordinary cameras and microphones, follows them against the person's own baseline, and hands the system a structured result it can act on.
With it, a system knows the place and the person, not just the place.
How the layer works is described on the Under the hood page.
First Live Run
On March 5, 2026, a live AI system received a real-time human-state input before generating a response.
Not from a questionnaire. Not from browser history. Not from behavioural assumptions.
From a live multimodal signal stream processed on the device.
Validation Statement
Current measurements are internal engineering measurements.
They do not constitute an independent third-party evaluation.
Published accuracy figures associated with underlying rPPG or facial-action methodologies belong to those individual methods and are not presented as performance benchmarks of EmoPulse.
Deployments
The same layer, in the places where it fits.
Driver Monitoring
The car knows the road. EmoPulse adds how the driver is doing.
Care & Senior Safety
Alarms answer an event. We are exploring whether slow changes in how a person is doing can be seen while help is still easy. Concept stage.
Clinical Workflows
A visit is a snapshot. We are exploring what happens between visits. Not a medical device.
Other Domains
Mobility, safety systems, AI platforms, robotics: wherever a system shares space with a person.
Part II: The Demonstrator
Demonstrating The Layer
The EmoPulse Dashboard is a public demonstrator of the layer working in real time.
The system computes human-state signals directly in the browser from a standard camera.
The public demos are one layer doing different jobs. What they share is the point: a system that already sees learns about the person too.
The demonstrator is not the product.
It is evidence that the layer operates in real time.
Part III: Intellectual Property & Privacy
Intellectual Property
Patented
The patented component is the architecture that turns signals into human state over time.
Open-source signal extraction components and published signal-processing methods are not claimed as proprietary inventions.
Privacy By Architecture
- On-device computation by default
- No video uploads
- No audio uploads
- No cloud dependency
- Air-gapped operation supported
- Optional server deployment
What Makes It Different
- One layer that adds the person to systems that already see the environment
- Human state over time, not a snapshot
- Several signals read together, so contradictions are information rather than noise
- On the device by default, so privacy does not depend on a promise
- Built on published methods and open components, with no dependence on a proprietary training corpus
- Built as a layer, not as an application
- Patented technology
The missing human-state layer
info@emopulse.app · emopulse.app
© 2026 EmoPulse. All rights reserved.
Patented technology.
Unauthorized reproduction, distribution, cloning, or use is prohibited. License