Straight answers.

No. Fire Watch is an assistive visual compliance layer that works alongside certified fire alarm systems. It never replaces statutory detection, alarms, suppression, or emergency procedures.

Not in Phase 1, by choice. Visual smoke/flame detection has high false-positive risk and is on our roadmap only after proper validation — and even then as advisory support, never a certified detector.

CCTV records; someone must watch it, usually after an incident. Fire Watch continuously evaluates defined fire-safety zones, raises advisory events with snapshot evidence, and drives them through verification, corrective action, and reporting.

We do not quote a lab figure. Accuracy is site-specific — it depends on camera placement and lighting. A pilot measures detection on staged violations, false alarms per camera per day, and time-to-closure, and reports those numbers to you.

Temporal validation (an obstruction must persist), cooldown suppression, conservative per-zone thresholds tuned during calibration, and a mandatory human review with a documented reason on every dismissal.

The system flags it — on the dashboard and in the daily report — and suppresses rules that need vision. It never hides its own blind spots.

Phase 1 runs on-premise with no cloud dependency. It stores event metadata and snapshots, not continuous recordings; access is role-based; there is no facial recognition.

Yes — that is the founding constraint. Standard RTSP/IP cameras, fixed views, 720p or better. No new sensors. A pilot begins with a camera-readiness survey.

Your safety team. Every event enters as advisory; an operator acknowledges or dismisses it with a reason, and a safety officer assigns and closes corrective actions.

A 2–4 week pilot on one site with 3–5 existing cameras, measuring detection, false alarms, and closure. Your certified fire systems continue unchanged throughout.