Carten ESB
The integration backbone. Connects your TOS, customs gateway, billing engine, and IoT equipment through a managed event mesh, so every system speaks one canonical language.
Carten is a modular platform that connects every actor in your port (terminal operators, customs, clearing agents, shipping lines, and authority staff) through a single managed event mesh. Six modules, including built-in AI intelligence. Run one on its own, or deploy the full platform.
A working terminal isn't one application. It's a TOS, a customs gateway, a billing engine, a fleet of equipment, dozens of agent firms, and a stream of vessels that never quite arrive on time.
Carten doesn't replace your TOS. It connects everything. A managed event mesh sits at the centre. Six purpose-built modules sit on top. All of them are optional and any of them runs on its own; they just share one event mesh, so adding the next one costs nothing.
The result is a port that runs on real-time data and AI intelligence built directly into the platform, instead of phone calls and spreadsheets.
Every module is independently deployable and licensed. They share one identity layer, one event mesh, and one design language, so data captured by one module shows up in the others without any integration work.
The integration backbone. Connects your TOS, customs gateway, billing engine, and IoT equipment through a managed event mesh, so every system speaks one canonical language.
Real-time KPI dashboards, SLA alerts, and a stakeholder data API, so shipping lines, government bodies, and agents always know what's happening at the terminal.
Mobile-first platform for the trade community: clearing agents, shipping agents, and freight forwarders. Declaration lodging, document management, customs status, and QR gate passes.
Mobile workforce operations for terminal staff: gate clerks, yard supervisors, and equipment operators. Shift management, incident reporting, equipment tracking, and digital task assignment.
The unified back-office. Single IAM for every module, system health monitoring, audit logs, agent ID card generation, and operational notice broadcasting, all in one place.
Platform-embedded AI, not a separate product. NL queries surface Pulse data in plain language. Document intelligence auto-extracts manifests. Predictive models flag delays and congestion before they happen.
Start with the module you need now. The others plug into the same mesh whenever you're ready for them.
Carten AI is not a separate tool; it's a platform capability surfaced inside the modules you already use. Ask Pulse a question in plain English. Upload a manifest and watch it become structured ESB data. Get vessel delay warnings before operations feel them.
All AI runs inside the terminal's own deployment namespace. Sensitive data (manifests, bills of lading) is processed on a local model. Terminal data never leaves the namespace without explicit approval.
Ask Pulse anything in plain English. The AI translates your question into a live query and returns an answer with data provenance, surfaced in the Pulse query bar. No SQL needed.
Upload a manifest, B/L, or customs declaration. Structured data lands in ESB store in seconds, with confidence scores and human-review flags for fields below the threshold.
Vessel delay forecasts, gate congestion warnings, and equipment fault risk surface as Pulse alerts and Worker task flags before they become incidents on the ground.
Conversational Agent
Deploy Carten AI as a native agent on the channels your port stakeholders already work in. No new app. No new login. The same live port data, surfaced in conversation.
Conversations are processed inside your terminal’s namespace. No data leaves your deployment.
A managed event mesh sits at the centre of every Carten deployment. Modules and external systems publish and consume against a canonical topic tree (port/{portCode}/{domain}/{entityId}/{eventType}), so adding a new system never breaks an existing one, and every event has a place to live.
Every integration is a publisher or a subscriber. Replace your TOS or swap your customs gateway; nothing else needs to change.
The mesh keeps a durable log. Replay events into any new module, analytics warehouse, or partner system without touching the source.
Canonical topic tree. Versioned schemas. Dead-letter queues. Audit trails on every message. Built for the long run, not the next demo.
Each role gets its own purpose-built tool: an SLA dashboard for the executive team, a rugged handheld for the gate clerk scanning QR passes, a phone app for the clearing agent tracking a consignment.
One pane of glass for vessel calls, berth occupancy, throughput, and SLA performance across every terminal, in real time.
Lodge declarations, generate invoices, pay duties, and collect cargo on a mobile-first platform that works in the queue at the gate.
Pre-arrival manifests, real-time vessel status, automated berth notifications, and stakeholder data feeds, all over the same event mesh.
Single-window declaration ingestion. Risk-routed clearance lanes. Live status broadcast to ICUMS, GCNet, ASYCUDA, or any national gateway.
Field-grade mobile app for shift logs, incident reports, equipment fault tracking, and digital task assignment. Works offline, syncs when it can.
Purple Engineering delivered the service digitization and system interoperability GPHA had been working toward for years. A competent and dependable partner — from first integration to post-deployment support.
Carten ships as containers. Each module deploys independently, with isolated state and clean upgrade paths. Self-host or fully managed by Purple Engineering.
We map your current systems (TOS, customs, billing, equipment) and identify which modules deliver value first.
Carten ESB and Carten Admin go in. The event mesh comes alive. Adapters to your TOS and customs are wired up.
One operational module (usually Pulse for visibility, or Port for the trade community) goes live in pilot.
Roll out remaining modules at your pace. Train your operators. We run managed services as long as you need.
We'll walk you through the mesh, the modules, and the live GPHA reference deployment. 45 minutes. No slideware, just a real port system running.