Content moderation with Jev

1. Pattern

• noul — does this violate policy X? (probability)
• choice — category (spam, harassment, ok, …)
• score — severity for queue priority

2. Thresholds in your code

Auto-allow when noul is clearly low, auto-block when clearly high, and send the middle band to humans. That is the operational win: you control cutoffs, Jev supplies calibrated probabilities.

3. Try it

Browse hosted tools for moderation-style demos, or call /api/v1/decide with your own policy questions.