We're betting plants want safety and operations on one camera tree.

The bet

Industrial plants have CCTV, smart phones, and operators who already log everything that matters. The hard part isn't the cameras or the AI — it's the workflow that turns a detection into a fix, and the rollout that earns each new duty rather than dumping it on the floor on day one. That workflow is what we're building.

Why now

Vision models that used to need a dedicated edge GPU now run on commodity boxes. Phones are powerful enough to act as portable sensors and incident-capture tools. And the EHS / operations leaders we talk to are tired of buying point solutions that don't talk to each other. The pieces line up.

Who we are

A small industrial-ops and applied-AI team based in Bengaluru. The first design partner is a plastic injection-moulding plant ten kilometres from where we write the code, which keeps the feedback loop short and the abstractions honest.

How we work

• Design partners, not customers. The first plants get the product at cost in exchange for honest feedback and the right to talk publicly when ready.
• Run quietly first. A new capability ships hidden until the data shows it earns its keep on one operator, then one shift, then the floor.
• Aggregate-only on operator-level signals. We don't score individuals. Trends drive equipment changes; people don't get scored by an AI.
• Video stays at the plant. The cloud sees metadata and selected stills, not the stream.

Where we're going

The next 12 months: deepen the safety + ops coverage that Wave 1 plants already see, ship the companion-app onboarding flow that lets a new plant come online without an integrator, and start publishing pilot results from our design partners. If any of that sounds useful for your plant, tell us.