Interkey products
SquintPRO is Interkey’s computer-vision platform for analysing CCTV streams, detecting operational events and generating alerts, evidence and KPIs. It runs on the cameras an organisation already owns, and it is available now.
Orientation
The whole page in one table, before the detail.
| Product | SquintPRO, an Interkey product, available now |
|---|---|
| Category | Computer vision, video analytics and operational intelligence |
| Capability | Configurable detections, cross-camera analysis, text and image event search, incident records and operational KPIs |
| Cameras | Works with existing CCTV estates; compatible cameras are not replaced |
| Deployment | On-premise, cloud or hybrid; GPU-based analytics scaled by camera stream |
| Processing location | Deployable in-Kingdom, including fully on-premise for sites that cannot send footage out |
| Object detection accuracy | 96.7% under the product’s reference conditions (see the note below: validate on your own footage) |
| Primary applications | HSSE monitoring (PPE compliance, restricted zones, unsafe proximity), automated inspection, operations and facilities monitoring |
| Alert delivery | Into the site’s existing VMS, alerting and incident workflow, not a new screen |
| Sectors deployed into | Oil and gas, ports and logistics, industrial manufacturing, utility networks, telecom operators |
| Delivery | Interkey owns the product and does the survey, integration, configuration and tuning in the Kingdom |
| Commercial model | Quote only: priced per site, camera count and use case |
The product
SquintPRO is an Interkey product, available today
SquintPRO is Interkey’s computer-vision platform for analysing CCTV streams, detecting operational events and generating alerts, evidence and operational KPIs. Interkey owns the product, builds it, and deploys and operates it on Saudi sites, so the software roadmap and the site engineering sit with the same company. The product was previously called Insight AI; the current name is SquintPRO. Its product site is squintpro.ai. This page is about applying it in the Kingdom: the problems it is bought to solve, what it does, how it deploys, and how to evaluate it.
The starting point
A large industrial site in the Kingdom typically has hundreds of cameras and a control room with a handful of screens. The footage is recorded diligently and watched almost never. When something goes wrong, it is reviewed afterwards. The cameras function as an archive rather than as a monitoring system.
Computer vision changes what that estate is for. Instead of a person watching for an event that may not happen, software watches continuously and raises an alert when a defined condition occurs: someone entering an area without the required protective equipment, a vehicle where vehicles should not be, a queue past a threshold, a piece of equipment in a state it should not be in.
The practical constraint is that this has to work on the cameras the organisation already owns. A proposal that begins by replacing a functioning CCTV estate is usually a proposal that does not proceed. SquintPRO adds the analytics layer to existing CCTV streams: compatible cameras stay where they are, and the estate starts producing events, evidence and numbers instead of only recordings.
Capabilities
Four groups of capability, in the order a site tends to use them: detect something, get someone to act on it, find it again afterwards, and turn the whole stream of events into numbers a manager can run a site by.
Detection categories are configured rather than coded: a site defines the conditions that matter to it, the zones they apply in, and the rules around them, through the interface. Custom detections can be built for cases specific to a site, and fine-tuned against samples from its own cameras. What is supported still depends on what is visible to the camera and what the model can be trained on, which the site survey establishes rather than assumes.
A detection raises a notification in real time and opens an incident record with the video evidence attached, so the event can be reviewed, assigned and closed rather than only noticed. Alerts are routed into the tools the site already runs, which is what keeps them from being ignored.
Analytics run across many camera streams at once, including analysis that follows subjects across cameras rather than treating each view in isolation. Past events can be found by describing them in text or by supplying an image, which turns an archive that was only ever searched by timestamp into something an investigator can actually query.
Detections accumulate into automated metrics and KPIs on customisable dashboards, with reports that can be shaped to how the site already reports. This is the part that outlasts the initial deployment: a continuous record of conditions by zone, shift and area rather than a monthly anecdote.
On the 96.7% figure.
Applications
The centre of gravity is industrial safety: the sectors these deployments run in are oil and gas, ports and logistics, industrial manufacturing and utility networks, which closely matches the client base Interkey already works with across government, energy and logistics. The same platform also serves operations, inspection and crowd monitoring where a site already runs cameras, but safety is where the case is strongest and where Interkey focuses.
The most common starting point. PPE compliance covers hardhats, high-visibility clothing, gloves, eye protection, safety footwear and harnesses where working at height applies. Alongside it sit the situational cases: restricted and no-go zones, line-of-fire exposure, workers close to moving equipment, work under a suspended load, and the unsafe acts a site already logs by hand, such as a person falling, lying on the ground, or a safety control bypassed.
Visual checks currently done on a schedule by a person walking a route: equipment and infrastructure condition, production-line observation, object counting, and detecting when a line stops or a pattern changes. Continuous observation catches what a weekly walk-round misses, and automated visual inspection can apply a site’s own criteria to what passes and what does not.
Headcount, density and occupancy against defined thresholds, pedestrian flow and congestion, and abnormal movement such as counter-flow or sudden dispersal. In an industrial setting this is zone occupancy and muster reporting; in a public venue it is the same capability applied to crowd management.
From camera to response
The technology only matters if the last stage works. Everything in this pipeline is configured around how the site already responds to safety events, because a detection that lands in a screen nobody owns is a detection that did not happen.
Existing cameras
The CCTV estate the site already owns, surveyed for which views are usable
Video analytics
SquintPRO detection on GPU, processed in-Kingdom, on-premise where footage must not leave the site
Safety event
A defined condition: missing PPE, entry into a restricted zone, unsafe proximity to moving equipment
Alert and evidence
Delivered into the site’s existing VMS and incident workflow, with the video record attached
HSE response
A supervisor acts, and the event becomes a record the safety system can count
Interkey builds and tunes every stage of this pipeline on site. The alert stage is deliberately not a new dashboard: detections are routed into the VMS and incident workflow the site already runs.
Deployment
The analytics layer attaches to the camera streams that already exist. Where it runs is a decision about data control and scale, and all three shapes are supported.
Analytics run on hardware inside the site network, which is what sites choose when footage must not leave the perimeter at all. Capacity is sized by the number of camera streams under analysis, and scales by adding GPU capacity rather than by re-architecting.
Central operation across multiple sites, with capacity that follows demand rather than being provisioned per location. This is the sensible shape for organisations running many small sites, or where a central operations team owns the monitoring.
Processing local to each site where the video is sensitive, with detection records, dashboards and reporting centralised. The split is decided per data class rather than per site, which is usually what a group-level HSE function actually needs.
Every one of the above can be kept inside Saudi Arabia. For operators whose video cannot leave the country, or cannot leave the site, that is a deployment choice rather than a special arrangement.
Delivery
SquintPRO is the platform; this is the work that makes it perform on a real site. All of it happens in the Kingdom, with your HSE team in the loop from the survey onwards.
Working out which of the existing cameras are actually usable for analytics. Camera placement chosen for human review is often wrong for automated detection: angles too high, resolution too low at the distance that matters, or a view into direct sun for part of the day. This assessment sets realistic expectations before anything is promised.
Turning a site’s safety rules into configured detections: which conditions, in which zones, at which thresholds, and what counts as a violation rather than a normal working state. Where a case is specific to the site, a custom detection is built and fine-tuned against footage from its own cameras.
Detections are only useful if they reach someone who can act. That means integrating with the VMS already in place and with whatever the site uses for alerting and incident management, so an alert appears in an existing workflow rather than in a new screen nobody has been assigned to watch.
The period after go-live matters more than the deployment. An analytics system that generates too many false alerts gets ignored within weeks, and an ignored system is worse than none because it creates a false sense of coverage. Threshold tuning against real site conditions is part of the work, not an afterthought.
Data stays put
Industrial video is sensitive twice over: it shows the site, and it shows the people on it. It helps to be precise about what a deployment actually produces, because three different things are casually called “the data”: the raw camera stream, the model’s inference output, and the detection record (a timestamp, a camera, a zone, an event class). Each can be held to a different rule, and the architecture should say which lives where.
Deployments can run entirely in-Kingdom, including fully on-premise where footage must never leave the site network. Worker privacy is handled as a design input, not an afterthought: what is retained and for how long, who can see it, how the workforce and contractors are informed, and where masking is applied. Interkey sets those controls with the site’s HSE and legal owners before go-live, and the record-keeping rules follow the site’s own policies rather than a vendor default.
For the safety team
Safety teams run on observations: unsafe acts and unsafe conditions logged, investigated and counted into leading indicators. The practical weakness of that system is coverage, because a supervisor sees a fraction of a shift, and behaviour changes when the supervisor is present. Continuous detection changes the volume and the honesty of that data: PPE non-compliance, zone breaches and proximity events get recorded whether or not anyone was watching, at 3pm and at 3am alike.
The output is deliberately shaped like the records an HSE system already keeps, so detections can feed the observation and incident process the site already runs rather than a parallel one. What that produces over time is a leading-indicator dataset large enough to manage by: which zones, which shifts, which contractors, which conditions, rather than a monthly anecdote.
Quick answers
Common question
Usually, for most of the estate. Analytics places different demands on a camera than human review: resolution at the distance that matters, angle, lighting, and a view that is not into the sun for half the day. The site survey establishes which cameras are usable as-is, which need adjusting, and which views genuinely need a new camera, before anything is promised.
Common question
An alert goes to whoever the site decides should act: the control room, the area supervisor, the HSE duty phone, inside the tools the site already uses. The event is also stored as an incident record with the video evidence attached, timestamp, camera, zone, event type, so it can be investigated and counted, not just reacted to. Designing that routing is part of the deployment, and it is where most of the operational value lives.
Getting started
The sensible entry point is a single site, a defined set of cameras and one or two detection cases with a measurable outcome. That produces a number for detection performance under your own conditions, which is the only figure that should drive a wider rollout decision, including in preference to the reference figure above.
Saudi industrial conditions are part of the test, deliberately: dust on camera domes, midday glare and heat haze, and the realities of a working site are exactly what a pilot should expose. Interkey has written up how to scope a pilot that proves something, including the success criteria worth insisting on with any vendor.
Why Interkey
SquintPRO is Interkey’s own product, which changes what a customer is buying. A detection that does not work on your cameras is a product question and a delivery question at the same time, and both sit with the same company: there is no vendor to escalate to and no roadmap that belongs to someone else.
Interkey has operated in Saudi enterprise technology since 1999, working across the government, energy, telecommunications and logistics organisations named on its client page, and computer vision sits inside that engineering practice rather than apart from it. The honest boundary: that history evidences the company relationships and the delivery capability, and per-site SquintPRO references are shared in conversation, with the client’s permission, rather than printed here.
The deployment work also rarely travels alone. Wiring detections into a permit or incident system is systems integration work, and deciding where inference runs on site is an infrastructure and platform question: both are practices Interkey runs in-house, which is what makes the last stage of the pipeline, the part where an alert becomes a response, an engineering deliverable rather than a hope.
From the insights library
Cameras and the site, detection and accuracy, data and privacy, delivery and rollout, one question at a time.
Solution guideCamera suitability, placement, false positives, privacy and what an alert workflow looks like on a working Saudi site.
Industry guideHazard scenarios, classified-area constraints, permit-controlled work and contractor workforces.
Evaluation guideThe success criteria, measurement discipline and honest failure modes of computer-vision safety pilots.
Interkey product
Built, deployed and supported by Interkey
Runs on existing cameras
Established by survey, not assumed
Riyadh delivery team
Survey, integration, tuning and operation
In Saudi enterprise technology since 1999
CR 1010156897
Next step
Tell us the site type, the camera estate you already have and the PPE rules that matter most, and the reply comes from the team that will actually walk the site.
Or directly
+966-11-2180999 info@interkey.com.saTawuniya Towers, North Tower, 7th Floor, King Fahad Highway, Olaya, P.O. Box 56835, Riyadh 11564, Saudi Arabia
Elsewhere