We build sensing and edge AI systems that combine cameras, sensors and on-device intelligence to understand the physical world in real time — turning multiple signals into actionable information at the point where it is needed.
Multi-modal sensing
Edge AI on-device
Multi-signal fusion · real-time inference
Actionable intelligence
Detection · prediction · automation · integration
Edge
On-device AI
Multi-modal
Any analog or digital signal
Real-time
Sense · infer · respond
Adaptable
Built around the application
Conventional monitoring waits for a single value to cross a line — too late, and too blunt to tell a real event from noise. Our platform reads the pattern of an emerging failure instead, and warns while there's still time to act.
From signal to action
Capture
Cameras and sensors capture visual, thermal, acoustic, electrical, environmental and other real-world signals.
Understand
On-device AI combines and interprets signals in real time, directly at the edge.
Act
Insights, alerts and automation reach the people and systems that need them, in real time.
Signal over time
The result isn't a binary alarm. It's a probability of the event — graded, early, and continuously updated — so you're warned while there's still time to act, with far fewer false alarms from everyday noise.
Stopyr applies our sensing stack to battery energy storage — multi-modal monitoring that catches thermal runaway before conventional BMS alarms react. Working prototype, tested against real battery packs. Targeting BESS deployments from residential to grid-scale.
Trace-level H₂, CO, and VOC gas detection before temperature moves
Per-module runaway probability from on-device edge ML
Retrofit-ready — plugs into existing BMS, SCADA, and site alarms
H₂ concentration
12 ppm
↑ Rising · above baseline
System status
Early signature
18% runaway probability
BESS Container 3 · Rack A
Cell array status
48
Modules
1
Early signature
Module grid · B-series
Illustrative UI · working prototype
Because it operates on time-series signals, the platform adapts across domains — the model changes, the platform doesn't. Here's where it stands today, honestly.
Our product Stopyr catches runaway before conventional alarms react — multi-modal gas sensing with on-device ML, tested against real battery packs. Targeting BESS deployments from residential to grid-scale.
See StopyrWe're developing edge vision and condition monitoring for rotating equipment and process safety — degradation and defects caught before they cost you.
Explore solutionsWe're developing edge-sensing approaches for clinical monitoring — early-stage, and open to pilot partners. If you run a clinic or health service with a monitoring problem, we'd like to talk.
Discuss a pilotNot every sensing problem has a product. We build custom edge-AI sensing systems for industrial and business environments — bring us the problem, and we'll tell you honestly if we can solve it.
Get in touchWe help Unicity Factory teams with the embedded systems and edge AI side of their projects: sensors, electronics, firmware and on-device models.
Our office was set up a few hundred metres from UMFST's new George Emil Palade University Clinical Hospital, Romania's first smart hospital and the government's pilot for digital healthcare.
Technology partner
Unicity Factory 2026
Mentor
Embedded systems & edge AI
Open lab
Hands-on STM32 edge AI for UMFST students