← Back to Hub

LUMINA Architecture

Complete overview of deployment infrastructure, services, and resources.

All Systems Operational
System Design

Deployment Architecture

Modular on-demand GPU system with separate frontend and backend services.

🌐
luminafold.com
Static Dashboard
↓
⚙️
Orchestrator
API & GPU Control
↓
🖥️
Lambda Labs H100
On-Demand Compute
↔
💾
S3 Storage
Persistent Data
Infrastructure

Deployed Services

All production services and their configurations.

🎨

Dashboard

Static Site • Render.com
URL luminafold.com
Repository lumina-dashboard
Type HTML/CSS/JS
Status Live
⚙️

Orchestrator

Web Service • Render.com
URL orchestrator
Repository lumina-orchestrator
Framework FastAPI (Python)
Status Live
🖥️

GPU Compute

On-Demand • Lambda Labs
Instance Type H100 SXM5 80GB
Region us-tx-1 (Texas)
Cost $2.49/hour
Mode On-Demand
💾

S3 Storage

Object Storage • AWS
Bucket lumina-data-foldartists
Region us-east-1
Contents Models, Generations
Access Private
🌐

DNS

Domain • Squarespace
Domain luminafold.com
Registrar Squarespace Domains
CNAME Target lumina-dashboard.onrender.com
SSL Active
🔐

Authentication

JWT Tokens • In-App
Method JWT (24h expiry)
Users 5 accounts
Hashing Bcrypt
Storage localStorage + session
Economics

Cost Breakdown

On-demand architecture achieves 95% cost savings vs always-on GPU.

$0
Dashboard (Free Tier)
$0
Orchestrator (Free Tier)
$2.49/hr
GPU (When Running)
~$3/mo
S3 Storage
Source Code

GitHub Repositories

All code is version-controlled and auto-deploys on push to main.

📁

lumina-dashboard

Frontend Repository
GitHub View Repo
Branch main
📁

lumina-orchestrator

Backend Repository
GitHub View Repo
Branch main