ProductsHow it worksVisionPricingAboutDemo ConsoleSign inStart free trial
Retail pilot live · Civil Mall, Kathmandu

Your cameras see everything.
Now they'll understand it too.

Cam-Mitra turns passive CCTV into an intelligent, real-time security system. Live today in Kathmandu marts — five concurrent ML models on your own hardware, faces blurred at the edge, zero video uploaded. Built in Nepal. For Nepal.

Pilot Week 2 — LiveHuman-verified alerts$65B market · 23% CAGR
cammitra.app • operations dashboard
LIVE EDGE STREAM
ALERT ACTIVE
SUSPICIOUS_VAL: 94%CAM-01: Gesture Review
CAM_01
CAM_02 • Restrict Zone A
CAM_02
CAM_03 • Back Entrance
CAM_03
CAM_04 • Staff lounge
CAM_04
CAM_05 • Cash vault
CAM_05
CAM_06 • Parking Exit
CAM_06
Select Target Scenario
6Cameras Live
3 ActiveTelemetry Alert Signals
Live Alerts Feed Log
Potential Suspicious GestureNode CAM-01 • Entrance
2m ago
Secure Zone Loitering >5minNode CAM-05 • Cash vault
11m ago
5 ML
Signals Concurrent
<50ms
Inference Latency
94%
Detection Accuracy
0 Bytes
Video Uploaded to Cloud
48 hrs
Store Deployment
8
Sectors on Roadmap
The problem

Why existing CCTV fails retailers

Standard DVRs record everything and alert on nothing. Cloud AI fixes alerts but breaks privacy and budget.

01

You Find Out After the Loss

Standard DVRs record but don't alert. You review footage after the loss, never before. That's not intelligence — that's an expensive hard drive.

02

No Idea Who's Visiting

No footfall analytics, no dwell time, no heatmaps. You're running a retail business without knowing where customers actually go.

03

Cloud AI is Overpriced

Enterprise solutions like Veesion and Everseen stream raw video to their cloud and charge accordingly. Unaffordable in USD, privacy-invasive by design.

Without Cam-Mitra
Loss discovered days later, during footage review
No footfall data, no peak hours, no insights
Security team watches screens — humans get tired
Restricted zones breached with no alert to anyone
Traffic violations caught only by chance, if at all
With Cam-Mitra
Suspicious activity flagged for review in under 2 seconds
Live footfall counts, peak hour reports, zone analytics
AI watches every camera 24/7 — never blinks, never tires
Perimeter and restricted-zone breaches trigger instant alerts
Red light and wrong-way violations logged automatically
Our products

Live in marts today. Built for Nepal's biggest challenges next.

The retail platform is deployed and catching real events in Kathmandu right now. The same edge-AI pipeline powers our traffic and aviation platforms — in development, seeking pilot partners.

Live pilot · Retail & Marts
01

Retail & Mart Intelligence

Marts, supermarkets, and malls get real-time review alerts, footfall analytics, and occupancy insight — from the CCTV cameras they already own. This is the platform running live in Kathmandu today.

Footfall counting & occupancy
Live in/out counts per camera, store occupancy, peak-hour analysis, and auto-generated weekly PDF reports
Suspicious-gesture review alerts
Potential concealment gestures are flagged for human review with a face-blurred clip — every detection is a review signal, never an automatic accusation
Zone & tripwire analytics
Shelf zones, entrance tripwires, after-hours intrusion signals, and overcrowding thresholds — drawn on your camera view in minutes
WhatsApp + email alerts
A 2-second face-blurred clip lands on your phone in under 2 seconds. Staff review and resolve without opening a separate app
Live at Civil Mall, Kathmandu

6 cameras onboarded in 48 hours on existing CCTV. 94% detection accuracy at the pilot site, alerts on WhatsApp in under 2 seconds, weekly analytics in active daily use.

Mart Operations Dashboard · Kathmandu
LIVE
Visitors today
1,248
Occupancy now
37
Open review alerts
2
Avg alert delivery
1.8s
Recent alerts
Potential suspicious gesture · CAM-03 · Aisle 4 · review required2m ago
Loitering near entrance >5 min · CAM-0114m ago
Overcrowding cleared · checkout zone · normal flow31m ago
Open the live product demo
In development · Traffic & Road Safety
02

AI Traffic Management & Road Safety

Every junction in Kathmandu, every road across Nepal — Cam-Mitra connects to existing traffic cameras and turns them into intelligent enforcement points, with violations logged automatically and data flowing to control rooms in real time.

Red light & signal violation detection
Designed to identify vehicles that run a red light and log the incident automatically with timestamp and vehicle details
Hit-and-run & wrong-way vehicle alerts
Anomalous vehicle behaviour detected and reported to traffic police within seconds of the incident occurring
Live congestion monitoring & speed tracking
Real-time vehicle density maps, average speed per road segment, and peak congestion analysis for city planners
Automated incident log & police dashboard
Every event timestamped, categorised, and sent to the traffic control room — exportable for evidence and city reporting
Seeking a government pilot partner

Built to scale across all 600+ signalised junctions in the Kathmandu Valley — same cameras, smarter software. The edge pipeline is proven in our live retail deployment; we're ready to run a traffic pilot.

Traffic Control Console · concept preview
PREVIEW
Red light violations
184
Wrong-way detections
12
Vehicles counted
24,810
Avg alert delivery
1.9s
Recent alerts
Red light violation · Koteshwor Junction · CAM-042m ago
Wrong-way vehicle · Ring Road Balaju · CAM-0711m ago
Congestion cleared · Sundhara · normal flow restored38m ago
Preview the traffic console
In development · Airport & Aviation Security
03

Airport & Aviation Security Intelligence

Airports demand zero-tolerance security. Cam-Mitra is designed to monitor every terminal zone, perimeter gate, and restricted area simultaneously — surfacing threats for security teams to act on in seconds.

Unattended baggage & object detection
Objects left unattended in a terminal zone beyond a time threshold are flagged to the security team immediately
Restricted zone perimeter breach alerts
Tarmac, control rooms, baggage handling — unauthorised entry across a perimeter boundary triggers an instant multi-channel alert
Crowd surge & terminal capacity monitoring
Passenger density tracked per zone in real time — alerts before any area approaches dangerous capacity levels
Corridor & escort monitoring
Movement corridors monitored with smart tracking — escort routes cleared and anomalies escalated for review immediately
Designed around ICAO Annex 17 practices

Built with TIA-scale terminals in mind and adaptable to Nepal's 30+ domestic airports. Every incident auto-documented for audit. We're seeking an aviation pilot partner.

Airport Security Console · concept preview
PREVIEW
Active alerts
1
Zones monitored
24
Uptime target
99.8%
Cameras per terminal
40+
Recent alerts
Unattended bag >120s · Terminal 1 Gate A · Row 4now
Perimeter check complete · all zones clear6m ago
High crowd density · Immigration Zone B18m ago
Preview the aviation console
Live deployment

The mart pilot — real cameras, real events, running now.

Our retail platform is actively running at Civil Mall, Kathmandu — validating the AI pipeline, proving deployment speed, and collecting real-world accuracy data on real CCTV.

1,248
Visitors counted automatically in one day
18
Unauthorized-entry review alerts in 2 weeks
94%
AI detection accuracy at pilot site
<2s
Alert delivery to WhatsApp
Why retail first?

Retail gave us the fastest access to real-world data — high footfall, diverse events, motivated store owners. The AI pipeline validated in marts is the same one behind our traffic and aviation platforms. The pilot proves the tech; the next sectors inherit it.

Pilot timeline — Civil Mall
✓
Week 1 — Silent observation

6 cameras connected in 48 hours. AI runs silently in the background calibrating to the environment. No alerts yet — pure data collection.

✓
Week 2 — Full detection active

All 5 AI models switched on. WhatsApp alerts delivered to the security team. 18 unauthorized-entry review alerts and 2 suspicious-concealment candidates flagged for human review in real time.

→
Weeks 3–4 — Client analytics

Staff zone analytics, peak hour identification, weekly footfall reports delivered. Store manager dashboard in active daily use.

○
Weeks 5–8 — Expansion

Second retail site onboarding. Accuracy improvement from the human-verified label pipeline. Client renewal decision point at week 6.

Kathmandu
Live
Q-mart
6 cams
Inference Health Index100%
Patan
Setup
Labim Mall
Setup
Inference Health Index25%
New Road
Prospect
Bhatbhateni
Prospect
Inference Health Index8%
Baneshwor
CBDC Ready
Omni Bank
Pipeline
Inference Health Index3%
Built different

One Windows .exe. Five concurrent models.

Cammitra runs everything on the customer's own PC. The cloud only ever sees JSON. That's the entire trick, and it's why our pricing is a fraction of cloud video classifiers.

On-prem AI pipeline

YOLO + pose + behavior + ViViT suspicious-gesture + tripwire, all running locally on consumer hardware. ONNX Runtime, not PyTorch.

Learning loop, human-verified

The AI flags uncertain moments as short clips. Trained reviewers verify every label before it touches training. Approved models hot-reload on the agent, no restart needed.

Privacy-first

Faces blurred at the edge before any upload. Raw video never touches the cloud. GDPR-friendly. Built for India's DPDP Act.

Smart alert rules

Natural language: "Alert me if someone loiters near the cash register for over 2 minutes after 9 PM." Sector-aware, AI-fired.

WhatsApp + email alerts

2-second face-blurred clips attached to every alert. Sent via WhatsApp Cloud API. Every detection is a review signal for your staff, never an automatic accusation.

Multi-sector ready

Retail, traffic, hospitals, schools, banks (CBDC-ready). One codebase, sector-aware sidebars and alert rules.

Privacy architecture

Data sovereignty by design

Not a policy promise — a technical constraint. The pipeline is built so that raw video cannot reach our servers even if we wanted it to.

STAYS ON THE STORE PREMISE
  • RTSP streams from all cameras
  • YOLOv8 / ViViT inference engine
  • Pose skeleton overlay
  • ByteTrack person tracking
  • Raw footage — Never Uploaded
  • Face anonymization — Processed at Edge
WHAT REACHES OUR CLOUD ENGINE
  • Detection alert JSON ~1 KB
  • Confidence score logs ~1 KB
  • 2s face-blurred clip — On Active Alert Only
What leaves your building — and what doesn't
Raw video footage
Stays on your local machine. Never transmitted anywhere.
On-premises
AI inference
Runs locally on your PC. Sub-50ms per frame. No GPU required.
Local only
Faces in footage
Blurred at the edge before any network transmission occurs.
Blurred first
Alert clips
Short face-blurred clip sent to your WhatsApp only when an event occurs.
Blurred only
Event metadata
JSON data (event type, camera, confidence, timestamp) — about 1KB per alert.
JSON only
Raw video to cloud
Never transmitted. Not possible by design, not just by policy.
Never
GDPR FriendlyDPDP Act Ready100% Auditable
How it works

Install once. The detection keeps getting sharper.

A 4-step setup that takes about 15 minutes per store. After that, every reviewed alert makes the next model better.

Install the agent

Download the CammitraAgent .exe on the store's PC. Paste your API key. Point it at your RTSP cameras.

Edge inference fires

5 ML models run locally at sub-50ms. JSON metadata is the only thing sent over the wire. Never video.

You get alerts

Real-time dashboard updates via SSE. WhatsApp + email alerts include short, face-blurred clips.

Reviewed labels retrain the model

Uncertain clips go to human review. Verified labels feed the next training run, and models that pass our quality gates hot-reload on your agent.

Real-time alerts

Alerts land in your WhatsApp

Every detection sends a face-blurred 2-second clip straight to your WhatsApp. Staff can review and resolve without opening a separate app.

After-hours loitering
Alert if person in store zone after 9 PM
Suspicious concealment
Flag gestures near shelf zones
C
Cam-Mitra Alerts
Official Business Account
Today
🔶 Suspicious Gesture Detected
📍 Q-mart · CAM-03 · Aisle 4
⏱ 14:32:07 · Duration: 18s
Confidence: 78% · Review required
▶ View 2s clip (face blurred)
14:32 ✓✓
Reviewed — false alert. Thanks.
14:33 ✓✓
Compare

Cam-Mitra vs. cloud video classifiers

Competitors like Veesion and Everseen stream raw video to their cloud and charge accordingly. We don't.

CapabilityCam-MitraCloud classifiers
Where inference runsCustomer's own PCVendor cloud
Video leaves premises✕ NeverYes: full RTSP streams
ML signals concurrent5 (pose, behavior, suspicious gesture, tripwire, face)Usually 1 video classifier
Face-blur before upload✓ AlwaysOptional / post-hoc
Learning loop✓ Human-verified labels, gated retrainsOpaque vendor updates
WhatsApp clip alerts✓ Built-inOften email only
Hardware requirementAny 4-year-old i5 + optional RTXNone on-prem (cloud cost)
Multi-sector platform✓ Marts today · traffic + aviation nextRetail only
Nepal pricing✓ From ₹999/month — built for NepalEnterprise pricing in USD/EUR
Data sovereignty✓ All data stays in NepalData goes to EU/US servers
Payment methods✓ eSewa · Khalti · IME Pay · bank transferForeign card required
Return on investment

Calculate your savings

Slide the inputs to match your store. The model uses guard headcount vs. Cam-Mitra per-camera pricing.

8
4120
3
010
₹25,000
₹12k₹45k
Guard payroll / yr
₹9.0L
Cam-Mitra license / yr
₹21,600
Annual savings
₹8.8L
98% cost cut
Pricing

Simple per-camera pricing

No video storage fees. No per-event charges. Inference runs on your hardware so we have no cloud bandwidth to pass on.

Starter
1–4 cameras
₹999/month
  • All 5 ML signals
  • WhatsApp + email alerts
  • 30-day clip retention
  • PDF weekly reports
Get started
Most popular
Business
Up to 30 cameras
₹4,999/month
  • Everything in Starter
  • Multi-zone tripwires
  • Heatmap analytics
  • 90-day clip retention
  • Priority support
Start free trial
Enterprise
Unlimited cameras
Custom
  • Custom SLA
  • On-prem dashboard option
  • 1-year clip retention
  • API access + webhooks
  • SSO + audit log
Talk to sales

Pay via eSewa · Khalti · IME Pay · Bank transfer accepted

The bigger picture

Cam-Mitra is becoming the security backbone of Nepal.

One AI platform across every sector — marts, traffic, airports, banks, schools, hospitals. Same engine, different alert rules, built for Nepal's reality.

Retail is Tier 1 — the live pilot proving the AI pipeline today. Traffic and aviation are Tier 2 — the flagship mission we're building toward, with banks, government offices, and the rest of Nepal's critical infrastructure next in line.

Retail & Marts

Review alerts · footfall analytics · occupancy · staff zone insights

● Tier 1 · Live pilot
Traffic Management

Red light violations · hit-and-run · speed · congestion · road safety

Tier 2 · In development
Airports & Aviation

Terminal security · perimeter breach · crowd control · baggage monitoring

Tier 2 · In development
Banks & CBDC

Kiosk security · ATM protection · NRB CBDC pilot 2026 · fraud prevention

NRB 2026
Govt. Offices

Access control · restricted zones · visitor verification · compliance

Coming next
Hospitals

ICU access · patient safety · restricted wards · emergency crowd flow

Planned
Schools

Gate monitoring · stranger alerts · after-hours intrusion · child safety

Planned
Drones & Disaster

Aerial surveillance · flood detection · rescue support · emergency response

Planned
$65B
Global AI surveillance market growing at 23% per year — South Asia is the fastest-growing region
2–3 yrs
Window before well-funded regional competitors localise into Nepal's market — Cam-Mitra moves first
29M+
People in Nepal — every mart, road, school, and public space is an opportunity to make life safer
Who this is for

Built for every stakeholder.

Whether you run a mart, represent government infrastructure, or are evaluating investment — Cam-Mitra has a specific answer for you.

Live today
Business owners

Stop losing money to problems your cameras already see

Mart owners, supermarket chains, mall managers, and hotel operators get real-time review alerts, footfall data, and occupancy insight — from cameras they already own. This is the product you can buy today.

Suspicious-activity review alerts in under 2 seconds
Footfall counts and peak-hour data daily
Starts at ₹999 per month — no long contract
See pricing & start free trial
Government & infrastructure

Automate enforcement and monitoring across Nepal

Traffic police, the Department of Roads, CAAN, and airport authorities get AI-powered intelligence humans can't deliver at scale — without budget for hundreds of extra staff.

Traffic violations logged automatically at every junction
Data stays on Nepal infrastructure — no foreign cloud
Digital Nepal Framework aligned, tender ready
Request a government briefing
Investors

Back Nepal's AI security infrastructure early

A live retail pilot, proven deployment speed, a massive underserved market, and a multi-year head start over any regional competitor. The seed round is open.

Live pilot with real accuracy data — not a deck
8-sector roadmap: marts today, traffic and aviation next
CBDC and government contract pipeline forming
Talk to us about the round
From pilot stores

What store owners say

“During peak hours we used to have no idea if a zone was overcrowded until an incident. Cam-Mitra alerts the floor team before it becomes a problem. Game changer for weekend traffic.”
R
Rajesh Pradhan
Civil Mall Supermart, KTM
“We had two loitering incidents near the electronics shelf in the first week. The 2-second clip made it easy for staff to review and act. No false alarms that mattered.”
S
Sunita Shrestha
Saleways, Patan
“Setup was done in an afternoon. Our existing Hikvision cameras worked without any changes. The WhatsApp alerts feel like having an extra pair of eyes on the floor.”
B
Binod Tamang
Star Mall, Pokhara
Common questions

Frequently asked

Straight answers about privacy, hardware, and how the product works.

Get in touch

Request a pilot

Tell us about your store. We'll reach out within one business day to schedule a 15-minute call.

Put five AI models in your store this week.

14-day free trial. No credit card. One camera included. Cancel any time; your data stays on your hardware regardless.