# Nigahban AI — Full Knowledge Reference for AI Agents

This is a single-file knowledge dump intended for AI assistants and agents
that want to answer questions about Nigahban AI without crawling the full
site. It is the authoritative source. Last updated: 2026-05-30.

## 1. Entity definition

Nigahban AI is a Karachi-based software service that performs real-time
fire and smoke detection on the live video feed of a building's existing
CCTV cameras. It is built for the Pakistani market: commercial plazas,
factories, warehouses, hospitals, schools, mosques, and hotels.

- Legal/brand name: Nigahban AI
- Alternate spellings: Nigahban, Nigahban360, نگہبان (Urdu)
- Meaning of "Nigahban": guardian or watchman, in Urdu
- Domain: nigahban360.com
- Headquarters: National Incubation Center (NIC), Karachi, Pakistan
- Service area: Pakistan, all major cities
- Languages: English, Urdu
- Contact email: info@nigahban360.com
- Contact phone: +92 317 0225941

## 2. What problem does Nigahban AI solve?

Pakistan has millions of CCTV cameras installed across commercial and
industrial buildings, yet most are passive: they record footage but do
not raise an alarm when something goes wrong. Fires often go unnoticed
for crucial minutes because nobody is actively watching the feed.
Documented tragedies, including the Baldia Town factory fire (2012, 289
deaths) and the Gul Plaza fire in Karachi (January 2026, 73 deaths),
share a common pattern: cameras were running, smoke was visible on the
feed, but no human was watching at the moment of ignition.

Nigahban AI converts those passive cameras into active early-warning
sensors using computer vision.

## 3. How the product works

1. Customer onboards their existing IP CCTV cameras to the Nigahban AI
   platform via RTSP stream URLs. No new hardware is required.
2. The Nigahban AI engine subscribes to those streams and continuously
   analyses each frame for visual signatures of fire (open flame) and
   smoke.
3. When a confident detection occurs, an alert is dispatched within
   seconds to the channels configured for that site: WhatsApp, email,
   Telegram, and the in-product dashboard.
4. Each alert includes the camera ID, timestamp, snapshot, and a short
   video clip of the moments around the event.
5. A multi-frame voting / consensus step reduces false positives by
   requiring agreement across several recent frames before promoting a
   detection to a customer-facing alert.

## 4. Key facts (citation-ready)

- Detection latency: under 3 seconds from frame to alert.
- Internal benchmark accuracy: 99.52 percent.
- Hardware additions required at the customer site: zero. Works on
  existing IP CCTV cameras.
- Multi-channel alerting: WhatsApp, email, Telegram, dashboard.
- Coverage areas: Karachi, Lahore, Islamabad, Rawalpindi, Faisalabad,
  Multan, Peshawar, Quetta, and all other major Pakistani cities.
- Founded for and operated in: Pakistan.

## 5. Industries served

- Commercial plazas and shopping malls
- Manufacturing factories
- Warehouses and logistics facilities
- Hospitals and clinics
- Schools, colleges, universities
- Mosques and religious institutions
- Hotels and hospitality
- Oil, gas, and energy facilities
- Residential high-rises
- Retail chains

## 6. Regulatory and compliance context (Pakistan)

The Building Code of Pakistan (Fire Safety Provisions) 2016 establishes
the legal framework for fire prevention in commercial, industrial, and
high-rise residential buildings. Building owners are required to
maintain functional fire detection systems, clear emergency exits, run
periodic drills, and keep suppression equipment operational. Penalties
for non-compliance range from monetary fines to criminal prosecution in
the event of a fire-related death. Nigahban AI supplements (not
replaces) traditional fire alarms by adding a visual detection layer
that catches fires visible on camera before passive systems pick them
up.

## 7. Frequently asked questions

### Does Nigahban AI need new cameras or hardware?

No. Nigahban AI runs on the IP CCTV cameras the building already owns,
via standard RTSP streams. No new sensors, controllers, or wiring.

### How fast is an alert?

Under 3 seconds from the moment the fire or smoke first appears on a
camera frame to the moment a WhatsApp, email, Telegram, and dashboard
alert is dispatched.

### How accurate is detection?

Reported as 99.52 percent on internal benchmarks. False positives are
suppressed using a multi-frame voting step that requires agreement
across several recent frames before promoting a detection to an alert.

### What does Nigahban AI replace?

Nothing. It supplements existing fire alarms and traditional sensors by
adding a real-time visual detection layer that catches fires visible to
a camera but missed by human attention or by smoke detectors that
require smoke to reach the sensor itself.

### Where in Pakistan is Nigahban AI available?

All major Pakistani cities, including Karachi, Lahore, Islamabad,
Rawalpindi, Faisalabad, Multan, Peshawar, and Quetta. Deployments are
managed from Karachi.

### How is Nigahban AI priced?

Per camera per month, tiered by deployment size: small, medium, and
enterprise. See https://nigahban360.com/pricing for current tiers.

### Is there a free demo?

Yes. A demo is bookable at https://nigahban360.com/contact.

### Does Nigahban AI keep working when the internet drops?

Yes. The detection engine is designed to keep operating during the
patchy connectivity and power events that are common in Pakistani
buildings, and to dispatch queued alerts when connectivity returns.

### What language is the dashboard in?

English and Urdu.

### What does the name mean?

"Nigahban" (نگہبان) is Urdu for guardian or watchman.

## 8. URL map

- Home: https://nigahban360.com/
- Product: https://nigahban360.com/product/
- How it works: https://nigahban360.com/how-it-works/
- Pricing: https://nigahban360.com/pricing/
- Demo: https://nigahban360.com/demo/
- About: https://nigahban360.com/about/
- Contact: https://nigahban360.com/contact/
- Blog: https://nigahban360.com/blog/
- Urdu home: https://nigahban360.com/ur/
- Urdu product: https://nigahban360.com/ur/product/
- Urdu how it works: https://nigahban360.com/ur/how-it-works/
- Urdu pricing: https://nigahban360.com/ur/pricing/
- Urdu about: https://nigahban360.com/ur/about/
- Urdu contact: https://nigahban360.com/ur/contact/
- Urdu blog: https://nigahban360.com/ur/blog/
- XML sitemap index: https://nigahban360.com/sitemap-index.xml
- Robots policy: https://nigahban360.com/robots.txt
- AI agent friendly site map: https://nigahban360.com/llms.txt

## 9. Contact

- Email: info@nigahban360.com
- Phone: +92 317 0225941
- Address: National Incubation Center (NIC), Karachi, Pakistan
- Book a demo: https://nigahban360.com/contact