Google Gemini 3 Hackathon Entry - Powered by Gemini 3.0
Globot represents a paradigm shift in global trade risk management. Unlike traditional static dashboards, Globot is an Agentic AI system that proactively monitors, analyzes, and mitigates supply chain risks in real-time.
π‘ "We are not just optimizing spreadsheets; we are securing the lifelines of the global economy."
When Globot helps Maersk reroute quickly during the Red Sea crisis, it means:
- πΎ Food supplies in Kenya remain uninterrupted
- π₯ Natural gas in Europe lasts through the winter
- π₯ Emergency hospital equipment doesn't get stuck at ports
This version (v2.1.20260131) is a specially built "High-Fidelity Mock" version, designed to ensure absolute stability and high impact during demonstrations. It simulates a realistic "4:55 PM Crisis Scenario" (geopolitical crisis in the Strait of Hormuz) and fully demonstrates the AI Chain-of-Thought and Human-in-the-Loop decision-making processes powered by Gemini 3.0.
- Real-time Reasoning Display: Shows the AI's thinking process line-by-line like a "typewriter," no longer a black box.
- Multi-Agent Debate: Features an "Adversarial Debate" (Red Team vs. Blue Team) to ensure decision robustness.
- Citations & Provenance: Every reasoning step is linked to specific documents in the RAG knowledge base or real-time news sources.
Features a team of 5 specialized AI Agents working in synergy:
- π Market Sentinel: Monitors geopolitical signals from Reuters/Bloomberg (Mock API supports multiple scenarios: Red Sea crisis, port congestion, etc.).
- π‘οΈ Financial Hedge Agent: Real-time analysis of fuel prices, exchange rates, and freight risks. Provides intelligent hedging strategies (futures, options, forwards) for both normal and crisis risk management. Dynamically calculates fuel costs (+$180K) and freight fluctuations after rerouting.
- π’ Logistics Orchestrator: Re-plans routes to avoid conflict zones.
- π Compliance Manager: Uses Gemini 3.0's Extended Context Window to analyze 500-page insurance policies and sanction lists.
- βοΈ Adversarial Debate: Red-teams decisions to prevent hallucinations.
Leverages Gemini 3.0 Vision multimodal capabilities:
- Satellite Image Analysis: Real-time detection of port congestion, canal blockages, and container accumulation.
- Suez Canal Scenario: Early warning for Ever Given-style events (6 hours before official announcements).
- Visual Evidence: Embeds satellite screenshots into decisions as reasoning evidence.
Showcases Gemini's long context window advantages:
- 500-page Maritime Insurance Policies: Automated parsing and route compliance verification.
- OFAC/UN Sanction Lists: Real-time checks (2M tokens context).
- MLC 2006 Convention: Automated verification of crew certifications.
- Routes & Ports: Visualizes major global shipping lanes and core ports like Shanghai, Rotterdam, and Los Angeles.
- Interactive Vessels: Click on yellow ship icons on the map to view detailed cargo manifests and voyage status.
- Dynamic Risk: When a crisis occurs, affected areas highlight and pulse with alerts.
- Decision Confirmation: After the AI makes a recommendation, a human must click "Approve & Execute" for it to take effect, reflecting responsible AI principles.
- Options: Provides "Details" and "Override" (manual intervention) options.
- Multi-channel Login: Integrated with Clerk, supporting Google, Facebook, LinkedIn social logins, and Email/SMS verification.
- Admin Console: A visual dashboard designed specifically for administrators to monitor global system KPIs.
- Security Whitelist: Email whitelist system based on environment variables to ensure privacy and security of administrative permissions.
The Reasoning Engine in this project is generic. Current use of Mock data is a Hackathon constraint. Production environments can directly integrate:
| Data Source | Purpose | Replacement Method |
|---|---|---|
| Bloomberg Terminal API | Real-time market data, geopolitical events | Replace mock_knowledge_base.py |
| MarineTraffic API | Real-time AIS vessel tracking | Replace demo/cot_data.py |
| Sentinel-2 Satellite API | Port/Canal satellite imagery | Replace visual_risk_service.py |
| Reuters/Bing News API | Real-time news feeds | Replace Market Sentinel data source |
| Industry | Example Companies | User Roles |
|---|---|---|
| Shipping & Logistics | Maersk, COSCO | NOC Manager, Control Tower Lead |
| High-end Manufacturing | Tesla, Apple | Global Supply Manager, Resiliency PM |
| Commodity Trading | Cargill, Glencore | Commodity Logistics Risk Lead |
| Freight Forwarding | Flexport | Trade Compliance Officer |
- Python 3.11
# Enter backend directory
cd backend
# Create virtual environment (recommended)
python -m venv venv
# Activate virtual environment
# Windows PowerShell:
.\venv\Scripts\Activate.ps1
# Windows CMD:
venv\Scripts\activate.bat
# macOS/Linux:
source venv/bin/activate
# Install dependencies
pip install -r requirements.txt
# Configure environment variables
# Create a .env file in the backend directory with core configurations:
# CLERK_ISSUER_URL=...
# ADMIN_WHITELIST=...
# Start server
python start_server.pyBackend runs at http://localhost:8000
cd frontend
npm install
# Configure environment variables
# Create a .env file in the frontend directory:
# VITE_CLERK_PUBLISHABLE_KEY=...
# VITE_ADMIN_WHITELIST=...
npm run devFrontend runs at http://localhost:5173
- Open browser and visit:
http://localhost:5173/pay - Click "Watch Demo" to go to the route selection page (
/port). - Confirm route (default Shanghai -> Rotterdam) and click "Start Simulation" to enter the demo (
/demo). - Observe the AI reasoning process and click confirm once the "Approve & Execute" button appears.
updates/README.md: Detailed demo operation steps (Must-read for new users).task.md: Project development task list.updates/README.md: Service startup and troubleshooting quick-reference table.
Globot now includes a comprehensive financial risk hedging system for managing:
- Fuel Price Risk: Hedge marine fuel price fluctuations using futures, options, and swaps.
- Currency Risk: Lock in exchange rates via forward contracts and currency swaps.
- Freight Rate Risk: Combined strategy of long-term charter contracts and spot market participation.
- β AI-powered risk assessment with Value at Risk (VaR) calculations
- β Automated hedging strategy recommendations (normal & crisis modes)
- β Real-time market data simulation
- β Crisis detection and emergency hedging protocols
- β Multi-instrument portfolio optimization
# Health check
GET http://localhost:8000/api/hedge/health
# Get market data
GET http://localhost:8000/api/hedge/market-data
# Assess risk exposure
POST http://localhost:8000/api/hedge/assess-risk
# Get hedging recommendations
POST http://localhost:8000/api/hedge/recommend
# Activate crisis hedging
POST http://localhost:8000/api/hedge/crisis-activate
# Generate executive report
POST http://localhost:8000/api/hedge/report- API Documentation:
backend/docs/HEDGING_API.md - Strategy Guide:
backend/docs/HEDGING_STRATEGY_GUIDE.md - Claude Skill:
backend/claude_skill/financial_hedging/SKILL.md
cd backend
python test_hedging_system.pyMaintainer: Vector897 License: MIT