"The same tool that helps you pick between Iceland and Singapore today will help you evaluate the lunar south pole tomorrow."
Skyly (Orbital Atlas) is a 3D interactive solar system explorer where users evaluate data center placement across Earth, the Moon, and Mars. Powered by AI-generated feasibility reports, blockchain-verified portfolio records, and real sustainability data — built in 48 hours at HackEurope Dublin 2026.
Rotate an interactive 3D globe, click any location — from Reykjavik to the lunar south pole — and get an AI-generated data center feasibility report covering cost, power, cooling, latency, carbon footprint, and risk. Then chat with an AI infrastructure advisor, build your portfolio, and get blockchain-verified receipts for approved sites.
Explore & Estimate — Rotate the 3D globe (Earth/Moon/Mars) → click a location → see a 6-axis scorecard (cost, power, cooling, latency, carbon, risk) → drill into a full construction blueprint
Ask the AI Advisor — Natural language queries like "I need 100MW with low carbon and <30ms to EU" → Claude analyzes 40+ locations against your existing inventory → recommends top sites with reasoning
Build & Track — Approve a recommended site → added to inventory → Solana receipt minted on-chain → Stripe payment processed → portfolio dashboard updates with real-time analytics
Portfolio Intelligence — Dashboard shows coverage maps, carbon footprint tracking, capacity metrics, and AI-powered gap analysis across your entire infrastructure
┌──────────────────────────────────────────────────────┐
│ FRONTEND (React + Three.js) │
│ ┌──────────┐ ┌──────────┐ ┌────────────────────┐ │
│ │ 3D Globe │ │ Chat │ │ Dashboard │ │
│ │ (Earth / │ │ Advisor │ │ (Inventory + │ │
│ │Moon/Mars)│ │ │ │ Portfolio Stats) │ │
│ └────┬─────┘ └────┬─────┘ └────────┬───────────┘ │
└───────┼──────────────┼─────────────────┼──────────────┘
│ │ │
▼ ▼ ▼
┌──────────────────────────────────────────────────────┐
│ BACKEND (Ruby on Rails) │
│ ┌──────────────┐ ┌───────────────────────────────┐ │
│ │ Location API │ │ Report Generator (Claude AI) │ │
│ │ (40+ sites) │ │ Scorecard + Blueprint + Chat │ │
│ └──────────────┘ └───────────────────────────────┘ │
│ ┌──────────────┐ ┌────────────┐ ┌──────────────┐ │
│ │ Stripe │ │ Solana │ │ Paid.ai │ │
│ │ Billing │ │ Minting │ │ Metering │ │
│ └──────────────┘ └────────────┘ └──────────────┘ │
└──────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────┐
│ SOLANA SERVICE (Rust + Axum) │
│ Memo transactions on Solana devnet │
│ On-chain portfolio records with tx verification │
└──────────────────────────────────────────────────────┘
| Layer | Technology | Purpose |
|---|---|---|
| Frontend | React 19, TypeScript, Vite | SPA framework |
| 3D Rendering | Three.js, React Three Fiber, react-globe.gl | Interactive Earth/Moon/Mars globes |
| UI | Tailwind CSS, shadcn/ui (Radix), Recharts | Components, charts, styling |
| Backend | Ruby on Rails 7.1, Puma | REST API server |
| AI | Claude API (Anthropic SDK) with structured tool use | Report generation, advisor chat, recommendations |
| Payments | Stripe Checkout + Webhooks | Per-blueprint billing ($299/report) |
| Blockchain | Solana devnet, Rust, Axum | On-chain portfolio record minting |
| Metering | Paid.ai | AI usage tracking and billing |
| Database | SQLite (dev) / PostgreSQL (prod) | Data persistence |
| Deployment | Docker, Docker Compose | Containerized services |
This project was designed to compete across 10 hackathon challenges simultaneously, integrating each sponsor's technology as a core feature:
| Challenge | Prize | How We Used It |
|---|---|---|
| Claude API | $10,000 | Core AI engine — generates all feasibility reports, powers the chat advisor, runs agentic tool-use loops for recommendations |
| Best Use of Data | €7,000 | Real energy prices, carbon intensity, climate data, latency metrics, and political stability scores across 40+ global and extraterrestrial locations |
| Crusoe Inference | $5,000 | Claude runs on Crusoe's clean-energy GPU infrastructure |
| Solana | €3,500 | Approved data center plans minted as verifiable on-chain portfolio records via memo transactions |
| Stripe | €3,000 | Per-report payment flow — free scorecards, paid blueprints with full Checkout + webhook integration |
| Rails | €2,000 | Entire backend API built on Ruby on Rails 7.1 |
| Sustainability | €1,000 | Core theme — evaluating environmental impact of data center placement globally and beyond |
| Lovable | €1,000 | Frontend scaffolded with Lovable, featuring 3D globe, dashboard, and landing page |
| Paid.ai | Office + 2yr | Usage-based metering tracks every AI report generated and value delivered |
| Adaptable Agent | Gift bags | AI adapts recommendations dynamically when inventory changes or energy prices shift |
| Consulting Agent | Team lunch | AI acts as an autonomous infrastructure consultant with tool access to portfolio data |
Total prize pool targeted: ~€30,500+
The backend uses Claude's structured tool use for agentic behavior. The AI advisor has access to 8 tools it can call autonomously in a loop:
| Tool | What It Does |
|---|---|
lookup_location |
Fetch detailed metrics for any location |
search_locations |
Filter locations by body, energy cost, carbon intensity |
compare_locations |
Side-by-side comparison of multiple sites |
calculate_costs |
Estimate build-out costs for a given capacity |
get_portfolio |
Read the user's current inventory and stats |
recommend_locations |
Analyze portfolio gaps and suggest optimal sites |
create_payment_link |
Generate Stripe checkout for blueprint purchase |
check_payment_status |
Verify payment completion |
- Scorecard (free) — 6-axis scoring (0-100), overall grade (A-F), executive summary, cost and timeline estimates
- Blueprint (paid, $299) — Phased construction plan, power strategy, cooling design, network topology, staffing plan, detailed cost breakdown, portfolio impact analysis
GET /api/locations # Browse 40+ curated locations
GET /api/locations/:id # Location detail
POST /api/reports/scorecard # Generate AI feasibility scorecard
POST /api/reports/blueprint # Generate full AI blueprint (paid)
POST /api/advisor/chat # Chat with AI advisor (tool use)
POST /api/advisor/recommend # Get AI recommendations
GET /api/inventory # User's data center portfolio
POST /api/inventory # Add data center to portfolio
PATCH /api/inventory/:id # Update data center
DELETE /api/inventory/:id # Remove data center
POST /api/payments/checkout # Create Stripe checkout session
GET /api/payments/session/:id # Check payment status
GET /api/payments/blueprints # List purchased blueprints
POST /api/webhooks/stripe # Stripe webhook handler
POST /api/solana/mint # Mint on-chain portfolio record
GET /api/portfolio/stats # Portfolio analytics
POST /api/predict/co2 # CO2 footprint prediction
GET /api/tle # Real-time satellite TLE data
GET /api/planetary/:body/:dataset # Moon/Mars region data
40+ curated locations across four celestial bodies, sourced from real-world data:
| Body | Locations | Data Sources |
|---|---|---|
| Earth | ~30 sites (Reykjavik, Singapore, Virginia, Dublin, Tasmania, ...) | IEA, Electricity Maps, NOAA, TeleGeography, World Bank |
| Moon | 5 sites (South Pole, Shackleton Crater, ...) | NASA Artemis, ESA lunar base studies |
| Mars | 3 sites (Hellas Basin, Olympus Mons, ...) | NASA Mars architecture studies |
| Orbit | 2 sites (LEO, GEO) | SpaceX rideshare pricing, satellite constellation data |
Each location includes: energy costs, carbon intensity, cooling methods, land costs, latency to markets (EU/US/APAC), disaster risk, political stability, regulatory environment, and connectivity options.
- Node.js 18+ and bun (frontend)
- Ruby 3.2+ and Bundler (backend)
- Rust and Cargo (Solana service)
- Docker (optional, for containerized setup)
Copy the example env files and add your API keys:
cp backend/.env.example backend/.env
cp frontend/.env.example frontend/.envRequired keys: CLAUDE_API_KEY, STRIPE_SECRET_KEY, STRIPE_WEBHOOK_SECRET, PAID_API_KEY
Frontend:
cd frontend
bun install
bun run dev # http://localhost:5173Backend:
cd backend
bundle install
rails db:create db:migrate
rails server # http://localhost:3000Solana Service:
cd solana
cargo run --bin keygen # Generate devnet wallet (first time)
cargo run --bin server # http://localhost:3001Docker (backend only):
cd backend
docker compose up --build # http://localhost:3002orbital-atlas/
├── frontend/ React 19 + Three.js + TypeScript
│ ├── src/
│ │ ├── pages/ Atlas (main app), Landing, Payment, Auth
│ │ ├── components/ GlobeView, ChatPanel, ScorecardPanel, InventoryPanel
│ │ └── lib/ API client, auth, constants
│ └── package.json
│
├── backend/ Ruby on Rails 7.1 API
│ ├── app/
│ │ ├── controllers/api/ 10 API controllers
│ │ ├── services/ Claude, Location, Paid, Prediction services
│ │ └── tools/ 8 Claude tool definitions
│ ├── Gemfile
│ └── Dockerfile
│
├── solana/ Rust microservice (Axum + Solana SDK v2.2)
│ ├── src/
│ │ ├── main.rs HTTP server (/health, /mint)
│ │ ├── mint.rs Memo transaction logic
│ │ └── wallet.rs Keypair management
│ └── Cargo.toml
│
└── docs/ Design docs and implementation plans
Built by a team of 4 at HackEurope Dublin 2026 (48-hour hackathon):
| Role | Responsibility |
|---|---|
| Person 1 — Frontend | 3D globe, UI panels, landing page (Lovable + React) |
| Person 2 — AI + Data | Location dataset curation, Claude prompts, report quality |
| Person 3 — Backend | Rails API, Claude integration, Stripe/Paid wiring |
| Person 4 — Blockchain + Integrations | Solana minting service (Rust), cross-service integration |
Built for HackEurope Dublin 2026.