Vengeea deploys on-premises AI video analytics at critical facilities across Singapore — intrusion, weapon, face, fire/smoke and license-plate detection running live on existing IP cameras, with theft and liquid-leak modules in active development. Because inference happens on the customer's own server inside Singapore, video and biometric data never leave the site, which is the architectural basis for PDPA alignment.
Vengeea runs on the customer's own infrastructure inside Singapore, so video and any biometric face-match templates are processed and stored on-site rather than routed to an overseas cloud region — which is what satisfies the PDPA's transfer-limitation obligation by removing the transfer itself, and aligns with IMDA and PDPC guidance for biometric and video-analytics deployments. That matters in practice at Singapore's density: a single vertical facility can carry hundreds of cameras across dozens of floors, and processing every one of those feeds on-site rather than backhauling it to a data center keeps both the compliance story and the network load manageable. Retention, deletion and role-based access are configured per site to fit the customer's DPIA.
All modules run in parallel on the same appliance, reading the same RTSP streams the site's cameras already output — no additional hardware, no camera swap. Below are the four modules Singapore's ports, industrial estates and high-density facilities request most, each live and generally available today.
AI intrusion detection — restricted-zone entry, line-crossing, after-hours access on existing IP cameras.
AI face analytics — watchlist matching against an on-prem face database.
AI weapon detection — firearm and edged-weapon detection in live video.
License plate recognition (LPR) — whitelist / blacklist alerts at gates and gantries.
Port of Singapore, financial infrastructure, industrial estates on Jurong Island, luxury retail, hospitality and public-sector critical infrastructure — dense, high-value sites where a security director needs both a defensible PDPA posture and detection that scales across camera counts without every additional feed adding to a cloud-transfer bill or a data-residency conversation.
Yes. Vengeea is deployed at critical facilities across Singapore under the on-premises model, running on the existing IP cameras already installed at each site. Because inference happens on the customer's own server, video stays inside the customer's own infrastructure inside Singapore, aligned with the PDPA's transfer limitation obligation without a cross-border-transfer question to answer.
Yes. Because processing happens on the customer's own hardware, Vengeea acts as data-intermediary with the customer as data-controller, and there is no cross-border transfer by default — the transfer-limitation obligation the PDPA imposes on any data leaving Singapore simply doesn't apply. Retention, deletion, access logs and role-based access are configurable per site to fit the customer's DPIA.
Any camera that outputs RTSP or ONVIF — Hikvision, Dahua, Axis, Bosch, Uniview, Hanwha and Milesight are all supported. There's no camera replacement, which matters in dense vertical facilities where re-cabling or swapping cameras is disruptive; integration is typically complete in 1-3 days from kickoff.
Yes. Vengeea deploys at ports and terminal operators at Port of Singapore scale; sizing is done per-site using the calculator, and because each site's appliance processes its own camera feeds locally, adding terminals doesn't multiply the amount of video routed off-site. Multi-site aggregation is optional and encrypted end-to-end.
1–3 day integration. No hardware replacement. On-prem by default.
