Grawizah
Team Leader & Full-Stack Architect
Overview
AI-native Global Trade Intelligence Platform empowering Indonesian MSMEs to scale globally. Built by team 'Successful Failures' (Wisnu as Team Leader, with Reza Fahlevi and Praisilia Pandoh from President University). Features neural AI matchmaking powered by Groq Llama 3.3, automated HS Code classification, real-time supply chain network mapping, competitor price benchmarking, and a multilingual trade translator supporting 14 languages. Won Web Developer category at TechSprint Innovation Cup 2026 organized by Codelab Indonesia.
The Problem
Indonesia's Micro, Small, and Medium Enterprises (MSMEs/UMKM) contribute more than 60% of the national GDP, yet fewer than 4% successfully enter the global export market. The root causes are structural and systemic: trust deficit in verifying international trade partners, information asymmetry where small suppliers have no access to real-time global market pricing or demand signals, regulatory complexity in HS Code classification and customs documentation, and language/communication barriers when negotiating with buyers from Germany, UAE, Japan, or Korea. Existing solutions like Alibaba, TradeIndo, and Tokopedia B2B address only listing and discovery — none provide integrated AI-powered intelligence, verification, and communication tooling in a single platform.
The Solution
Grawizah is an AI-native Trade Intelligence Platform that serves as a strategic command center for international trade — not just a marketplace. It features a neural AI matchmaking engine powered by Groq Llama 3.3 that analyzes production capacity, ISO certifications, trade history, and market demand signals to produce ranked buyer-supplier matches with human-readable rationale. An interactive Trade Network Map visualizes global supply chain relationships. The Neural HS Code Classifier automates customs classification with confidence scoring. A Buyer/Supplier Radar provides live-stream trade demand monitoring. Competitor Price Benchmarking delivers real-time pricing intelligence from multiple global sources. A Multilingual Trade Translator supports 14 languages for trade-specific terminology. The platform serves two distinct roles: Supplier Portal for Indonesian exporters and Buyer Portal for international procurement teams.
Key Features
AI Smart Matchmaker
Neural matching engine powered by Groq Llama 3.3 that analyzes production capacity, ISO certifications, trade history, and market demand signals to produce ranked buyer-supplier matches with human-readable rationale for every recommendation.
Interactive Trade Network Map
Real-time, interactive node graph visualizing global supply chain relationships, route dependencies, and market concentrations. Suppliers can identify at-risk routes and alternative market opportunities at a glance.
Neural HS Code Classifier
AI-powered classification of any product description into the correct international Harmonized System (HS) Code with confidence scoring, ensuring customs compliance from the start of the export process.
Buyer/Supplier Radar
Live-stream radar feed monitoring global trade demand signals, helping suppliers spot high-intent buyers and helping buyers discover emerging suppliers before competitors do.
Competitor Price Benchmarking
Real-time price intelligence pulled from multiple global sources, displayed with regional breakdowns and visual bar charts, enabling both buyers and suppliers to negotiate from a position of knowledge.
Multilingual Trade Translator
Embedded translation engine specialized for trade and procurement terminology, supporting 14 languages including Indonesian, Chinese, Arabic, Japanese, and Korean for seamless cross-border communication.
AI Listing Optimizer
Automated scoring and improvement recommendations for supplier product listings to maximize visibility and conversion on the platform, with actionable suggestions for optimization.
Inquiry & RFQ Management
Full-featured inquiry pipeline with status tracking (Open, Responded, Closed), conversion tracking, response time analytics, and buyer rating system — all role-aware and updated in real-time.
Architecture
Next.js 14 (App Router) frontend with TypeScript and Tailwind CSS communicates via REST API to Go (Golang) backend with high-concurrency request processing. JWT authentication middleware with bcrypt + RS256 provides stateless, role-aware security. Supabase (PostgreSQL) with Row Level Security provides real-time subscriptions and full-text search. Groq Llama 3.3 powers AI inference for matching, classification, and optimization. Rate limiting is handled in-process with per-role tier limits (Free/Premium).
How to Use
Register as a Supplier (exporter) or Buyer (importer) at the platform URL. Suppliers create product listings with detailed specifications, certifications, and pricing. The AI Smart Matchmaker automatically suggests potential buyers or suppliers. Use the Trade Network Map to visualize supply chain relationships and identify market opportunities. The Neural HS Code Classifier can classify any product description into the correct international HS code. Track inquiries through the RFQ management pipeline with status tracking and response time analytics. Use the AI Listing Optimizer to improve product listing visibility. The multilingual translator helps communicate with international partners in 14 languages.
Impact & Results
Grawizah was developed for the TechSprint Innovation Cup 2026 organized by Codelab under the theme "Smart Digital Solution for Real World Problems" and won the Web Developer category. The platform addresses the critical gap where fewer than 4% of Indonesian MSMEs can export by providing Fortune 500-level trade intelligence tools at zero marginal cost. The AI Smart Matchmaker powered by Groq Llama 3.3 reduces due diligence time from weeks to seconds. Real-time competitor benchmarking eliminates under-pricing and over-pricing risks. The platform was built by Team "Successful Failures" from President University with 3 members across Product Management, Frontend/UI-UX, and Full-Stack Engineering.