AI Infrastructure Platform with Managed API Service
A complete AI foundry platform for orchestrating data generation, model distillation, training, and evaluation on ephemeral GPU compute.
# 1. Start Backend API
export PYTHONPATH=$PYTHONPATH:$(pwd)/src
uvicorn nexa_compute.api.main:app --port 8000
# 2. Start Frontend Dashboard (new terminal)
cd frontend && npm run dev
# 3. Start Mintlify Docs (new terminal)
cd docs/mintlify && npx mintlify dev
# 4. Populate Demo Data
python scripts/create_dashboard_demo.pyNexa_compute/
├── src/nexa_compute/api/ # FastAPI backend
│ ├── main.py # Main application
│ ├── auth.py # API key authentication
│ ├── database.py # SQLAlchemy models
│ ├── endpoints/ # API routes
│ └── services/ # Business logic
├── frontend/ # Next.js dashboard
│ ├── app/ # App Router pages
│ ├── components/ # React components
│ └── lib/ # API client
├── sdk/ # Python SDK
│ ├── nexa_forge/ # Client library
│ ├── setup.py # Package config
│ └── demo.py # Demo script
├── docs/mintlify/ # Documentation
│ ├── mint.json # Mintlify config
│ ├── *.mdx # Doc pages
│ └── logo/ # Branding assets
└── scripts/ # Utility scripts
- ✅ 6 job types (generate, audit, distill, train, evaluate, deploy)
- ✅ Worker management & orchestration
- ✅ API key authentication (SHA256 hashed)
- ✅ Metered billing tracking
- ✅ Real-time job status & logs
- ✅ Real-time dashboard with metrics
- ✅ Expandable job logs
- ✅ Worker fleet monitoring
- ✅ Billing analytics with charts
- ✅ Secure API key management
- ✅ Dark theme UI
- ✅ Simple client API
- ✅ All 6 job types supported
- ✅ Environment variable config
- ✅ Comprehensive documentation
- ✅ Getting started guides
- ✅ API reference
- ✅ Architecture diagrams
- ✅ Pricing information
- ✅ SDK examples
| Feature | Free Tier | Pro Plan |
|---|---|---|
| GPU Hours/Month | 10 | Unlimited |
| Concurrent Jobs | 2 | 50 |
| Job Retention | 7 days | 90 days |
| Support | Community | Priority |
| SLA | None | 99.9% |
| Price | $0 | $99/mo + usage |
from nexa_forge import NexaForgeClient
# Initialize client
client = NexaForgeClient(api_key="nexa_...")
# Generate data
job1 = client.generate(domain="biology", num_samples=100)
# Train model
job2 = client.train(
model_id="llama-3-8b",
dataset_uri="s3://bucket/data.parquet",
epochs=3
)
# Monitor jobs
status = client.get_job(job1['job_id'])
all_jobs = client.list_jobs(limit=10)- API keys hashed with SHA256
- One-time key display on creation
- Secure modal with warnings
- Revocation support
- Ready for rate limiting
- Backend: FastAPI, SQLAlchemy, SQLite
- Frontend: Next.js 16, Tailwind CSS, Recharts
- SDK: Python 3.11+
- Docs: Mintlify
- Deployment: Docker Compose ready
./scripts/start_forge.sh- Frontend → Vercel
- Backend → Railway/Render
- Docs → Mintlify
- Database → PostgreSQL
Full documentation available at http://localhost:3001
Key sections:
- Getting Started: Quick setup guide
- Architecture: Platform overview
- API Reference: Complete endpoint docs
- SDK Guide: Python client usage
- Pricing: Freemium model details
See CONTRIBUTING.md for guidelines.
Built with ❤️ using FastAPI, Next.js, and Mintlify