Elemental Genius Unified Learning Platform
An AI-powered adaptive learning platform for children (K-12) with emphasis on early learners and special needs education. Built as cloud-native containerized microservices on Google Cloud Platform.
Owner: Elemental Genius LLC Status: Live on GCP — all services deployed to Cloud Run, frontend on Firebase Hosting
Prodigee follows a containerized microservice architecture deployed on GCP Cloud Run. AI, translation, speech, and vision capabilities are externalized to GCP APIs — no self-hosted inference.
Users → Firebase Hosting (prodigee-488119.web.app)
│
├── /** → Static Next.js SPA (CDN-cached)
│
└── /api/** ──────────────────────────────────┐
│
┌─────────────────────────────────────▼┐
│ API Gateway (Cloud Run) │
│ Routing, rate limiting, ID tokens │
└───┬────┬────┬────┬───────────────────┘
│ │ │ │
┌─────────────┘ │ │ └──────────────┐
│ │ │ │
┌──────▼──────┐ ┌───────▼──┐ ┌▼───────────┐ ┌───▼────┐
│ Auth │ │ Learning │ │ Analytics │ │ AR/VR │
│ JWT, COPPA │ │ Engine │ │ Progress │ │ Parked │
│ Sessions │ │ Curriculum│ │ Dashboards │ │ │
└─────────────┘ └──────────┘ └────────────┘ └────────┘
│
┌──────────┴──────────┐
│ GCP APIs │
│ Vertex AI │
│ Speech-to-Text │
│ Cloud Translation │
│ Cloud Vision │
└─────────────────────┘
Hosting: Firebase Hosting (CDN + Cloud Run rewrites)
Compute: Cloud Run (us-east1)
Data: Firestore (nam5 multi-region US)
Images: Artifact Registry (us-east1)
Secrets: Secret Manager
Auth: Service-to-service via metadata server ID tokens
See ARCHITECTURE.md for the full technical specification.
prodigee/
├── services/ # Microservices
│ ├── gateway/ # API Gateway — routing & rate limiting
│ ├── auth/ # Authentication — JWT, COPPA, sessions
│ ├── learning-engine/ # Curriculum delivery, Vertex AI, Speech
│ ├── analytics/ # Progress tracking, dashboards, reporting
│ └── ar-vr/ # AR/VR services (post-launch)
│
├── clients/ # Frontend applications
│ └── web/ # Next.js 14 + TypeScript + Tailwind CSS
│ └── src/app/ # Pages: dashboard, activity, progress, auth
│
├── tests/ # Integration tests
│ └── test_integration.py # 8-phase E2E flow against docker-compose stack
│
├── shared/ # Shared libraries
│ ├── models/ # Firestore document schemas (Pydantic)
│ ├── utils/ # Common utilities
│ └── config/ # Shared configuration
│
├── infra/ # Infrastructure & deployment
│ ├── setup-secrets.sh # GCP Secret Manager setup
│ ├── setup-domain.sh # Cloud Run domain mapping
│ └── terraform/ # GCP resource definitions (planned)
│
├── eg-curriculum/ # Curriculum data (JSON) — reusable as-is
├── eg-backend-code/ # [Legacy] Flask monolith — reference only
├── eg-android-app/ # [Legacy] React Native mobile app
├── eg-frontend-win-mac/ # [Legacy] Electron desktop app
├── eg-vr-classroom/ # [Legacy] VR classroom services
├── library-backend-data/ # [Legacy] FastAPI content server
├── library-indexer/ # [Legacy] Content indexing tool
├── eg-institute/ # [Legacy] Admin portal stub
├── _archive/ # Archived previous implementations
│
├── .firebaserc # Firebase project alias
├── firebase.json # Firebase Hosting + Firestore rules config
├── ARCHITECTURE.md # Master technical architecture spec
├── DEVELOPMENT.md # Local development setup guide
├── docker-compose.yml # Local dev orchestration
└── cloudbuild.yaml # GCP Cloud Build CI/CD pipeline (12 steps)
Folders prefixed with eg- and library- contain legacy code from the original build. This code is being decomposed and rewritten into the services/ microservices. The legacy folders remain for reference during the rewrite.
Do not build new features in the legacy folders.
| Service | Port | Purpose | Key GCP APIs |
|---|---|---|---|
| gateway | 8080 | Request routing, rate limiting, auth validation | — |
| auth | 8081 | Parent/child auth, COPPA compliance, JWT sessions | Identity Platform, Firestore |
| learning-engine | 8082 | Curriculum delivery, adaptive learning, voice | Vertex AI, Speech-to-Text, Translation, Storage |
| analytics | 8083 | Progress tracking, dashboards, teacher reporting | Firestore |
| ar-vr | 8084 | AR/VR interactive experiences (parked) | Vision |
The platform delivers three integrated curricula:
- Heggerty Phonemic Awareness — 35-week systematic progression across 8 skills (rhyming through phoneme substitution), 8-minute daily lessons with multi-sensory hand motions
- STEM Integration — Hands-on projects organized by age group (3-5, 6-8, 9-10)
- Character Development — 12 core traits with biblical foundations and American values
Curriculum data lives in eg-curriculum/ as JSON and is consumed by the learning engine service.
See DEVELOPMENT.md for full setup instructions.
# Clone
git clone git@github.com:ProfessorMoose74/prodigee.git
cd prodigee
# Run all services locally
docker compose up --build
# Services available at:
# Gateway: http://localhost:8080
# Auth: http://localhost:8081
# Learning Engine: http://localhost:8082
# Analytics: http://localhost:8083
# AR/VR: http://localhost:8084
# Firestore UI: http://localhost:8086| Property | Value |
|---|---|
| Public URL | https://prodigee-488119.web.app |
| Project ID | prodigee-488119 |
| Project Number | 446166630230 |
| Region | us-east1 |
| Firestore | nam5 (multi-region US) |
| Artifact Registry | us-east1-docker.pkg.dev/prodigee-488119/prodigee/ |
| Firebase Hosting | Serves static SPA + proxies /api/** to Cloud Run |
| CI/CD | Push to main → Cloud Build → deploy all services + hosting |
| Service | Cloud Run Name | Status |
|---|---|---|
| Gateway | prodigee-gateway |
Deployed (--no-invoker-iam-check for Firebase Hosting) |
| Auth | prodigee-auth |
Deployed |
| Learning Engine | prodigee-learning-engine |
Deployed |
| Analytics | prodigee-analytics |
Deployed |
| AR/VR | prodigee-ar-vr |
Deployed (stub) |
The gateway fetches Google ID tokens from the Cloud Run metadata server to authenticate with backend services. User JWTs are forwarded via the X-Forwarded-Authorization header. Backend services check this header first, falling back to Authorization for direct access.
- Backend: Python 3.12 + FastAPI + Uvicorn
- Frontend: Next.js 14 (App Router) + TypeScript + Tailwind CSS
- Data: Google Cloud Firestore
- AI: Vertex AI (adaptive learning, content generation)
- Speech: Cloud Speech-to-Text (pronunciation assessment)
- Translation: Cloud Translation API (30+ languages)
- Auth: JWT with COPPA-compliant child sessions
- Secrets: GCP Secret Manager
- Hosting: Firebase Hosting (CDN + Cloud Run rewrites)
- Containers: Docker → Cloud Run
- CI/CD: Cloud Build → Artifact Registry → Cloud Run + Firebase Hosting
- Testing: pytest + pytest-asyncio (unit) + httpx E2E integration
| Project | Status | Platform |
|---|---|---|
| TruckerFlow | Final stages | GCP |
| CreatorFlow | Pending MS Store cert | Windows |
| ServiceFlow | Live | Google Play |
| TidyFlow | Active | Desktop |