A company that lets users post content a forum, a social app, a marketplace, a game chat needs moderation. They sign up on Vireon, get an API key, and send their content to your endpoint. Sentra analyzes it and returns a verdict. They never have to build moderation themselves.
Their App Vireon
───────── ──────
User posts comment ───────► POST /v1/analyze/content
◄─────── { "verdict": "toxic", "confidence": 0.94, "latency_ms": 12 }
App blocks comment
| Layer | Technology |
|---|---|
| Backend | Python + FastAPI |
| Database | PostgreSQL + SQLAlchemy |
| Cache | Redis (rate limiting + result caching) |
| Local AI | HuggingFace transformers — unitary/toxic-bert |
| Auth | JWT for dashboard + X-Keys for API access |
| Billing | Stripe (usage-based) |
| Deployment | Fly.io |
This project was built to demonstrate a moderation pipeline not to be enterprise, It does not feature billing as intended.