Getting Started
An introduction to the customs compliance and trade facilitation platform.
What This Platform Does
This platform is a web-based customs compliance and trade facilitation tool designed for Trinidad and Tobago importers, exporters, freight forwarders, and customs brokers. It provides:
- HS code search and classification -- look up Harmonized System commodity codes with natural language queries, powered by semantic vector search
- Duty estimation -- calculate applicable duty rates across 11 rate types including ICD, VAT, surcharge, CCD, DRD, MVT, and CARICOM preferential rates
- Customs declaration preparation -- generate pre-populated C82 declaration data and ASYCUDA-compatible XML output
- Document intelligence -- extract invoice and shipping data from uploaded commercial documents using Azure AI
- Reporting and analytics -- track classification activity, duty estimates, and usage patterns
Who It Is For
The platform is built for the customs brokerage workflow in Trinidad and Tobago, specifically:
| Role | What They Do |
|---|---|
| Customs brokers | Classify goods, estimate duties, prepare declarations for filing |
| Importers/Exporters | Look up HS codes and duty rates before shipping |
| Freight forwarders | Estimate landed costs for clients |
| Compliance reviewers | Audit classifications before submission |
Current Status
The platform is deployed in production at sesko.netlco.com. Core features -- HS code search, duty calculation, AI classification, document extraction, and ASYCUDA XML export -- are functional and have been tested against real commercial scenarios.
The tariff data is sourced from publicly available information. For production-grade accuracy, the following are needed:
- Official tariff data from the T&T Customs and Excise Division (to supplement publicly sourced data)
- Official ASYCUDA XML specifications from UNCTAD (to validate the reverse-engineered format)
- Validation against production filing outcomes with practicing customs brokers
Technical Foundation
The platform is built on modern cloud infrastructure:
- Runtime: Next.js (React) on Azure Container Apps
- Database: PostgreSQL 16 with pgvector for semantic search
- AI Services: Azure OpenAI (GPT-4.1-mini for classification, text-embedding-3-small for search) and Azure Document Intelligence for OCR
- Authentication: NextAuth v5 with role-based access control (6 roles)
- Infrastructure: Microsoft Azure (East US), Cloudflare CDN/WAF
See Platform Overview for full architecture details.