Babi Product Intelligence System

Automated scraping system for extracting product ingredient data and related policy/tariff information for Babi's beverage brands.

Products

• KOH Coconut - 100% coconut water from Thailand
• VIVALOE - Aloe Vera drinks with real Aloe Vera
• NOYU Teas - Functional teas and bubble tea

Quick Start

cd scraper
pip install -r requirements.txt

# Set your OpenAI API key
$env:OPENAI_API_KEY = "your-key-here"

# Run the scraper (LLM auto-extracts config from Brand.md)
python main.py --brand-file ../Brand.md

Results saved to scraper/output/

See scraper/README.md for full documentation.

Scraper Workflow

Phase 0: LLM Configuration (Automatic)

🤖 GPT analyzes Brand.md and extracts:

• Brand names (KOH Coconut, VIVALOE, NOYU Teas)
• Website URLs (all sites per brand)
• Target countries (current + expansion markets)
• Product descriptions

No manual configuration needed! Just update Brand.md.

Phase 1: Product Data Scraping

🕷️ Scrapes all configured websites:

• https://www.kohcoconut.com/home-usa
• https://www.kohcoconutworldwide.com
• https://www.vivaloeworldwide.com
• https://www.noyuteasworldwide.com

Extracts: Product names, ingredients, descriptions, volumes

Phase 2: Policy & Tariff Data

📊 Based on extracted ingredients, searches:

• WTO ePing - Food safety regulations (SPS/TBT notifications)
• World Bank WITS - Global tariff data
• WTO Tariff Database - Official tariff schedules

For markets: Automatically uses target countries from Brand.md

Output Files

• product_config.json - LLM-extracted configuration (cached)
• products.json / products.csv - Product ingredient data
• policy_notifications.csv - Food safety regulations
• tariffs.csv - Tariff rates by country
• policy_data.json - Complete policy dataset

Documentation

• scraper/README.md - Complete scraper documentation
• Brand.md - Babi's products and markets

Key Features

Ingredient Monitoring

• Tracks restrictions on coconut water, aloe vera, tea ingredients
• Detects banned/restricted substances (e.g., aloin in aloe vera)
• Monitors preservative and additive regulations

Tariff Analysis

• Collects MFN rates, preferential rates, FTA rates
• Covers priority markets: USA, Canada, Europe
• Tracks VAT and excise taxes

Regulatory Alerts

• SPS (food safety) notifications
• TBT (technical barriers) notifications
• Comment deadlines for new regulations
• Compliance requirements by country

Product HS Codes

Product	HS Code	Description
KOH Coconut	2009.89	Coconut water, not concentrated
VIVALOE	2202.99	Aloe vera beverages, non-alcoholic
NOYU Teas	2202.10	Tea-based beverages, ready to drink

Target Markets

High Priority

• United States - Sacramento warehouse, direct distribution
• Canada - Existing VIVALOE market (Montreal), expansion target

Medium Priority

• Ireland - Existing KOH market
• United Kingdom, Germany, France - European expansion

Expansion

• China, Hong Kong - Existing KOH markets
• Cambodia - Existing VIVALOE market
• Guam - Existing KOH market

Project Structure

startup/
├── scraper/
│   ├── main.py              # Main scraper (use this!)
│   ├── requirements.txt     # Dependencies
│   ├── README.md           # Full documentation
│   └── output/             # Generated data
├── Brand.md                # Product information
├── README.md               # This file
└── (legacy files for reference)

License

Internal use only - Babi Corporation

Road Map

• Phase 1 ( Foundations & Backend Setup )

  • Design Database Schema (user, field, product, country, policy/tariff)

• Phase 2 ( Initial Data Ingestion )

  • Develop a targeted web scraper
  • Build an internal manual data entry form

• Phase 3 ( The Dashboard MVP )

  • Implement simple user login/authentication
  • Create a basic dashboard UI
  • Display all companies in a clean, filterable table