Automate. Validate. Ship with Confidence.
TradeCheck is an AI-powered documentary validation platform built for foreign trade operations. It replaces the manual, error-prone process of cross-checking commercial documents with an intelligent, standardized, and professional system that generates audit-ready reports in seconds.
Foreign trade operations depend on the accuracy and consistency of multiple documents — Commercial Invoices, Packing Lists, and Bills of Lading. A single discrepancy can trigger customs retention, clearance delays, or financial penalties.
TradeCheck solves this by acting as a Senior Foreign Trade Documentation Specialist powered by AI. The platform:
- Accepts uploads of the three core trade documents
- Extracts structured data from each file
- Cross-references all fields across documents
- Classifies every inconsistency by risk level
- Generates a professional validation report — shareable and exportable
The MVP is designed for internal validation with a single client, replacing a manual ChatGPT-based workflow with a scalable, repeatable system.
| Feature | Description |
|---|---|
| 📄 Multi-document Upload | Upload Invoice, Packing List, and BL in PDF or DOCX |
| 🤖 AI Extraction | Structured field extraction per document using GPT |
| 🔍 Cross-document Validation | Automatic comparison of shared fields across all documents |
| Every inconsistency tagged as Low / Medium / High risk | |
| 📊 Confidence Score | AI transparency score for each analysis |
| 📝 Structured Report | Executive summary, detailed findings, and final recommendation |
| 🔗 Shareable URL | Public report link at /share/{analysis-id} — no login required |
| 📥 DOCX Export | Download the report as a formatted Word document |
| 🖨️ Print View | Browser-optimized printable report layout |
| 🕓 Analysis History | Last 50 analyses stored locally and reopenable |
- React + TypeScript
- TailwindCSS — desktop-first, enterprise UI
- React Router — SPA navigation and shareable report routes
- Node.js + Express
- REST API for document processing, analysis orchestration, and report management
- OpenAI API — GPT-4o / latest available model
- Prompt-engineered to behave as a Foreign Trade Documentation Specialist
- PDF extraction —
pdf-parseor equivalent - DOCX extraction —
mammothor equivalent
- Local JSON storage (MVP)
- Architecture prepared for PostgreSQL migration
┌─────────────────────────────────────────────────────────────┐
│ FRONTEND (React) │
│ │
│ Upload Center → Validation Panel → Report Viewer │
└──────────────────────────┬──────────────────────────────────┘
│ REST API
┌──────────────────────────▼──────────────────────────────────┐
│ BACKEND (Express) │
│ │
│ /upload → /extract → /analyze → /report → /share │
└──────────────────────────┬──────────────────────────────────┘
│
┌────────────────┴────────────────┐
│ │
┌─────────▼──────────┐ ┌──────────▼─────────┐
│ File Processor │ │ OpenAI API │
│ (PDF / DOCX) │ │ (GPT Model) │
└────────────────────┘ └────────────────────┘
│
┌──────────────────────────▼──────────────────────────────────┐
│ Local Storage (JSON) │
│ [ Prepared for PostgreSQL migration ] │
└─────────────────────────────────────────────────────────────┘
- Node.js 18+
- npm or yarn
- OpenAI API Key
TradeCheck supports three core foreign trade documents:
Commercial Invoice
- Supplier, Exporter, Importer
- Invoice Number, Invoice Date
- Product Description, HS Code
- Quantity, Unit Value, Total Value
- Incoterm, Gross Weight, Net Weight
Packing List
- Supplier, Importer
- Gross Weight, Net Weight
- Quantity, Number of Packages, Volume
- Container Number
Bill of Lading (BL)
- Shipper, Consignee, Notify Party
- Container Number, BL Number
- Gross Weight, Vessel
- Port of Loading, Port of Discharge
The AI compares shared fields across all three documents and flags:
- Supplier / Shipper mismatches — different names across documents
- Weight discrepancies — gross/net weight inconsistencies
- Quantity mismatches — units differ between Invoice and Packing List
- Container number conflicts — BL vs. Packing List container IDs
- HS Code issues — missing or conflicting tariff classifications
- Incoterm inconsistencies — missing or contradicting trade terms
- Incomplete documents — mandatory fields left blank
| Level | Meaning |
|---|---|
| 🟢 Low Risk | Minor or cosmetic inconsistency, unlikely to cause operational issues |
| 🟡 Medium Risk | Relevant inconsistency that should be corrected before shipment |
| 🔴 High Risk | Critical inconsistency with direct customs or financial risk |
Every analysis produces a structured report with four sections:
- Overall documentation status
- Global risk level
- General recommendation
For each issue identified:
- Problem description
- Location in the document
- Risk classification
- Potential operational impact
- Suggested correction
- Fields that are consistent and correctly filled across all documents
- Overall assessment
- Final recommendation before shipment
Every report includes an AI confidence score (e.g., 94% — High Confidence) based on the completeness and readability of the uploaded documents.
After every analysis, a unique public URL is generated automatically:
https://yourdomain.com/share/{analysis-id}
The shareable page displays the full report — Executive Summary, Detailed Analysis, Positive Findings, and Conclusion — with no login required. This allows freight forwarders, customs brokers, and clients to access reports directly.
Note: Report URLs are public by design for the MVP. Authentication and access control are planned for a future version.
The MVP architecture is built to support the following future features without requiring a full rewrite:
- User authentication and role-based access
- PostgreSQL database migration
- Multi-company / multi-tenancy support
- Audit logs
- Custom validation rule sets per client
- Knowledge base for trade regulations
- Learning system based on historical analyses
- External API integrations (customs systems, ERP)
- Workflow automation and CRM integrations
| Token | Value |
|---|---|
| Primary | #FF6B00 (Orange) |
| Background | #111111 (Black) |
| Surface | #2A2A2A (Dark Gray) |
| Neutral | #F5F5F5 (Light Gray) |
The UI is designed to look and feel like enterprise logistics software — professional, executive, and internationally focused.
This project is proprietary software developed for internal client validation. All rights reserved.