Most businesses already own everything they need. The point of sale, the cameras, the door, the WhatsApp thread, the payment rail, the accounts. None of it talks. Lynexis builds the software that is missing and integrates the rest, so the work moves without anyone re-typing it.
We are not a reseller and not a body shop. Every engagement is scoped so a specific manual process disappears at the end of it, and we stay on the hook once it does.
Web applications, internal tools, customer-facing storefronts and the data model underneath them. Shipped to production, not to a slide.
Wiring the platforms a business already pays for into one another so records stay in step and nobody copies data between screens.
Facial recognition at the door, the gate and the desk, integrated with the roster, the tenant list and the audit trail behind it.
The physical layer a terminal or a camera sits on. Power to the door, network to the gate, cable that carries the next hundred devices.
Agents and event-driven workflows that read, decide and act, with a human check on anything that touches money or a customer.
Someone answers when a gate stops reading or a nightly job fails. Systems we put in are systems we keep running.
Storefronts, operations platforms and marketplaces, all on infrastructure we run ourselves. Where money moves, it settles straight into the merchant's own account with no intermediary holding it.
Ordering storefront with PayNow SGQR checkout, per-order bill reference and direct bank settlement.
Second tenant on the same codebase, themed independently and deployed without touching another merchant's data.
Social commerce marketplace on a falling-price clock, with per-seller SGQR settlement and a Telegram broadcast layer.
Events and catering operations. Enquiry to quote to work order to invoice to profit and loss, in one pipeline.
Booking and operations for licensed travel agencies, built around group departures and instalment collection.
First merchant on our bounce-back wallet. Diners claim credit from a receipt and redeem it on the next visit, with money going straight to the merchant.
Events automation across mailbox, task board and work orders, plus a direct ordering platform, built in-house across a three-outlet group.
Tell us which two systems refuse to talk to each other and we will map the integration before anyone writes code.
A full facial recognition stack we deploy, integrate and support in Singapore. Detection, landmark mapping, attribute classification, liveness and one-to-many search, available as an embedded toolkit, an on-premise platform or a finished hardware terminal.
Capture and enhancement pipeline that produces usable enrolment images from poor light and low-resolution sources.
Locates and tracks faces in live video and stills across angle, occlusion and crowd density.
Feature-point mapping for alignment, pose correction and downstream matching accuracy.
Classification of visible attributes for analytics, segmentation and operational reporting.
One-to-one verification against a held record, or one-to-many search across a population database.
Anti-spoofing that separates a live subject from a photograph, a screen or a recorded video.
Matches a live face against the photograph on an identity card, at both low and high resolution.
Licensed modules for embedding detection, capture and matching directly into your own application.
Wall-mounted and desktop units that verify a person at the door using face, fingerprint, identity card or any combination of the three. Standalone, or wired into an access control panel you already own.
Turnstile and speed-gate integration for high-throughput lobbies, transit points and ticketed venues, where identity and entitlement are checked in the same pass.
Server-side systems for organisations managing identity at scale, including large population search and recognition layered over an existing camera network.
For teams building their own product. Licensed modules that drop the recognition engine into your application, on server, desktop or mobile.
Configured application platforms for the environments where checking who someone is holds up everything behind it. Each one is fitted to the site rather than sold as a box.
Verifies candidates at the hall entrance and again at the desk, closing the gap that manual document checks leave open.
Binds each station's results to a verified identity, so records follow the person through the screening without paper forms.
Tenant, staff and visitor access across lobbies, lifts and restricted floors, with a single audit trail.
Party verification and role-based entry for proceedings and restricted facilities, enforced at each threshold.
Contactless clock-in that cannot be passed to a colleague, reconciled against roster and payroll.
Identity document matched to a live face with liveness checking, for check-in and account opening flows.
At a controlled point, a doorway, a gate, a check-in desk, the conditions are fixed and the subject is facing the lens, so a match can be trusted. An open floor is a different problem. Grease on the lens, kitchen smoke, a delivery bike parked at the entrance, and the numbers stop being credible. So out there we invert it. The camera fires an action. The systems that are already exact, the till, the invoice, the booking, record the result.
Door contacts, point of sale event streams and power monitoring corroborate what the camera sees, so a blurred week does not break the record.
Inference runs on a site box. Only events leave the network, never footage, which keeps the deployment inside PDPA expectations.
Alerts arrive on WhatsApp the moment they matter, with one monthly statement reconciling what the cameras saw against what the accounts recorded.
Everything below is operated in house, not rented from a vendor and marked up. It is the same stack we deploy for clients.
Lynexis was founded by Iman Nordin, who spent years as a business owner running and engineering operations across events, security and logistics before building software full time.
That order matters. The problems we solve are the ones we used to have at seven in the morning, with a job starting at noon and three systems disagreeing about the same number.
We work with owner-run businesses in Singapore and the region. Small teams, real operations, no appetite for a twelve month implementation.
Engagements start with a paid scoping session that produces a written integration map you keep, whether or not you build it with us.