Solutions

Video analytics solutions for the cameras you already own.

Quadinel AI adds an intelligence layer to the IP cameras and VMS you already run. Every video analytics capability works on the same edge hardware and reports to the same console, so the next one is a licence, not another rollout.

DETECTIONS RUNNING IN LIVE NETWORKS

Gas leak · unit 4

Forklift zone · warehouse

Container ID · port gate

Stock aisle · shelf gap

Footfall · mall entrance

Site activity · zone 3

Queue length · counter

Pothole · lane 2

Crowd density · concourse

Gloves & vest · line 3

Weapon detected

Heat map · store aisle

Drone inspection · plant

Fire & smoke

Cycle time · cell 4

School zone · safe entry

Illegal parking

Wrong way

Vehicle count · corridor

Gas leak · unit 4

Forklift zone · warehouse

Container ID · port gate

Stock aisle · shelf gap

Footfall · mall entrance

Site activity · zone 3

Queue length · counter

Pothole · lane 2

Crowd density · concourse

Gloves & vest · line 3

Weapon detected

Heat map · store aisle

Drone inspection · plant

Fire & smoke

Cycle time · cell 4

School zone · safe entry

Illegal parking

Wrong way

Vehicle count · corridor

Fewer accidents. Faster response. Lower cost per camera.

Nobody buys AI for its own sake. Operations teams want fewer accidents, tighter compliance, incidents handled while they are still small, less lost to theft and error, and more done in every shift. That is what our video analytical solutions are measured against, on the camera network you already run and without a cloud bill for every stream.

Catch it while it is happening

A blocked fire exit, a forklift in a pedestrian lane, a person down on a platform. The alert reaches your control room in the same second as the frame, not in the next morning's incident report.

ALERTS · LAST HOUR

No hard hat, zone B 09:41

Overspeed, lane 3 09:38

Coolant pooling, bay 3 09:22

Show the evidence behind every number

PPE rates, near misses, queue times. Each figure carries the clip and timestamp that produced it, so a safety review or an audit gets footage rather than an estimate.

EVIDENCE TRAIL

Emission event #1284 09:41:07

Clip attached 14 s

Signed off by HSE lead

Run more on what you already have

One deployment carries every capability you add to it. Models are sized for edge hardware, there are no per-stream fees, and a new use case does not mean a new box in the server room.

PER-CAMERA COST

Cloud, per stream 100%

Quadinel, on the edge 46% (figure to be confirmed by the client)

Caption: Models 50 to 80% smaller, no per-stream fees, no extra box per use case.

One primary solution. Three services alongside it.

Rotating labels on the traffic image. Replace the Indian plate format with a Saudi

PLATE · 1234 ABC / VEHICLE · SEDAN / SPEED · 58 KM/H

HEAVY · LANE 3 / VEHICLE · TRUCK / ATCC · CLASS 04

01 . VIDEO ANALYTICS · PRIMARY SOLUTION

Intelligence for the cameras you already have

Detection models run on an edge appliance beside your recorder, reading the RTSP streams your IP cameras already publish to the VMS. No new cameras per use case, no footage leaving the site, and one console for traffic, safety, retail, logistics and security applications. This is the solution every industry section on the homepage is built on.

200+ models Runs on your premises Custom classes in weeks

02 · ESG REPORTING PLATFORM · SEPARATE SERVICE

ESG numbers your auditor can trace to the source

Most sustainability reports in the Kingdom are still assembled from spreadsheets, utility bills and emails a few weeks before the deadline. Our ESG reporting platform replaces that with one ledger. Energy, fuel, water, waste, emissions, safety and workforce data flow in from meters, ERP, fleet systems and site logs, each entry keeping the document, reading or approval that produced it. The platform maps that data to the frameworks Saudi companies actually report against: the Saudi Exchange ESG disclosure guidelines, GRI, and IFRS S1 and S2, with greenhouse gas accounting across Scope 1, 2 and 3 under the GHG Protocol. When the assurance provider asks where a figure came from, the answer is one click away, not a week of reconstruction.

Tadawul, GRI and IFRS S2 Scope 1, 2 and 3 Assurance-ready evidence

03 · CYBERSECURITY · SEPARATE SERVICE

Security you can show a regulator, not just a dashboard

Saudi organisations now answer to a specific set of controls: NCA ECC-2:2024 for government and critical infrastructure, NCNICC-1:2025 for the wider private sector, SAMA CSF for financial institutions, and the PDPL for personal data. Our cybersecurity service starts with a gap assessment against the framework that applies to you, so the work that follows is tied to a control you have to evidence, not a generic checklist. From there we test and fix. Vulnerability assessment and penetration testing across web, mobile, API, network, cloud and OT. Network segmentation, hardening, identity and access controls, patch and credential management. Then continuous monitoring, with alerts triaged and escalated, and an incident response plan that meets the breach notification windows SDAIA and SAMA expect.

NCA ECC and NCNICC VAPT and hardening Continuous monitoring

04 · DIGITAL AND AI SERVICES · SEPARATE SERVICE

Powering Smarter Government with Digital & AI.

Quadinel AI helps government organizations accelerate digital transformation through practical AI and technology solutions. We support organizations in improving public services, automating processes, connecting data and systems, optimizing resources, and enabling faster, smarter decision-making. Our approach focuses on scalable solutions that integrate with existing environments while supporting innovation, operational efficiency, and the development of future-ready government services.

Arabic voice agents Document intelligence Process automation

Built on Awiros, deployed by Quadinel AI PLATFORM & SOFTWARE

Every video analytics capability on this page runs on the Awiros video AI platform and its appliances. We size the hardware for your camera count, licence the applications your operation needs, connect the console to your VMS and run it with your team. What you end up with is a working deployment, not a component list.

PLATFORM & SOFTWARE

UVIP

Platform

Unified video intelligence platform

The console every capability reports into. It takes feeds from cameras, recorders and third-party VMS, runs inference in real time and routes alerts to the people who act on them, with role-based access for the control room, HSE and management.

Multi-source ingest Real-time alerts Role-based access

AppStack

Application library

120+ vision AI applications

The catalogue our capabilities are licensed from: detection classes across mobility, safety, retail and industry. Adding a capability is a licence on the appliances you already run, not new hardware.

120+ apps Per-site licence Days to add

aiscope

Handheld

Inspection where there is no camera

A handheld unit for field and plant inspection. Automated scanning, anomaly detection and instant alerts, with models tuned to the asset being checked.

Portable Anomaly detection Field inspection

ATOM

Annotation

Training data for the classes you need

Precision annotation for detection, segmentation and tracking. It is the tool behind a custom class when your use case is not already in the library.

Detection Segmentation Tracking

HYDRA

Cloud testing

Check accuracy before hardware is sized

An evaluation environment for confirming what each application reads before an appliance is specified. Testing is set up within your own environment and under your control, in line with the Kingdom's rules on camera recordings.

Under your control Accuracy report Before hardware

Appliances

V-Stream Enterprise

2U rack

Network-scale inference

A rack-mountable appliance for large deployments, preloaded with the platform OS and running 150+ deep-learning applications from one box in your server room.

150+ apps Rack-mounted Enterprise

V-Tower

Tower

Mid-size sites, plug and play

Turns an existing CCTV network into an analytics system without a server room. Sized for a single site or a mid-scale network of cameras.

Plug and play Single site CCTV retrofit

EdgePlus-7

Edge

Remote locations, no backhaul

Built for highways, warehouses and remote retail. Inference happens on site with very little bandwidth to carry, while the console manages every node from one place.

On-site inference Low bandwidth Central console

Awiros Cloud

Cloud

Where on-premise is not required

The same platform on public or private cloud, including in-Kingdom hosting where data residency applies. Scaled per feed rather than per box.

Public or private Per-feed scaling In-Kingdom hosting

Deployment models

On-premise

Everything stays inside your infrastructure. Full control of footage and data for operators under strict privacy or regulatory mandates.

Edge and cloud

Inference at the edge, management in the cloud. The lowest latency, with a fraction of the bandwidth to carry.

Cloud

No appliances to host, scaled per feed. The quickest to stand up when hardware on site is not an option.

Questions we hear before a deployment

No. The platform is intelligence for the cameras you already have. Detection runs on an edge appliance beside your existing recorder, against the RTSP streams your IP cameras already publish, and alerts can surface inside the VMS your operators use today. An ONVIF-compatible feed at a reasonable resolution is enough. Most deployments need no new cameras at all.

On your premises, on your hardware. Inference happens on the appliance inside your network. Only the detection result is written to the console: a timestamp, a class, a bounding box and, if you choose, a still. That console lives with you as well. Nothing is sent to Quadinel AI and we never ask for recordings, which keeps a deployment within the Kingdom's Security Surveillance Cameras Law and the PDPL, both of which restrict where camera recordings and personal data can be transferred.

Around 100 use cases can be deployed in 7 working days, with minimal resources from your side. The applications are already built and licensed from the library, so the week goes into mapping cameras to use cases, setting alert rules and routing them to the right people, not into development. What we need from you is network access, a place for the appliance and someone who knows the camera layout. Larger networks scale the same way, appliance by appliance.

The library ships with 200+ models, from perimeter intrusion to queue length, and new classes are trained and delivered in weeks. Walk us through what you need detected, on site, and we will tell you plainly whether it is a training job, a camera-placement problem or something the platform is not built for.

Number plate recognition is trained per region. For the Kingdom that means bilingual plates with Arabic and Latin characters, read at highway speed, at night and through dust. Accuracy is validated on your own roads during the pilot, before any enforcement use is discussed.

No. They are separate services with their own scope, team and agreement. You can engage any of them without deploying video analytics, and a video analytics deployment does not include them. We list them on one page because the same clients often need more than one, not because they share a platform.

Video analytics software is licensed per capability, per site, on annual terms. Edge hardware is quoted separately or supplied by you. Because appliances are shared across capabilities, the cost per camera falls as you add more. The additional services are scoped and quoted on their own.

Any IP camera that publishes an RTSP or ONVIF stream, and any recorder or VMS that can pass those streams on. The Riyadh retail deployment ran on a standard Hikvision NVR with no changes. Recorded video on the VMS can be analysed as well as live streams.

Yes. Results can be viewed in the platform console, inside your VMS client, or in your command and control centre, and pushed as alerts and data to your reporting tools. Your operators do not have to learn a new screen unless they want to.

Each site gets its own appliance and processes its own video. Only counts, events and snapshot evidence travel over your existing WAN to head office, where a consolidated view compares and ranks sites. Bandwidth stays minimal and data residency does not change as the estate grows.

Every deployment starts with a short observation period. Queue zones, dwell times, tolerance windows and sensitivity are tuned on live conditions, and a false-alert review is part of the first two weeks. After go-live there is a monthly accuracy and performance review, and thresholds stay yours to adjust.

Access is role-based. A site manager sees their own views, head office sees the consolidated estate, engineers see device health. Each user's view is limited to what their role needs, and every event carries the camera and timestamp it came from.

A baseline week of clean data, a user manual in two parts, training for site and head office teams, agreed alert routing and escalation, then scheduled reports and a monthly review. New use cases are added from the library as licences, not as projects.