Ask clearly
Jev receives one focused question for profanity, slurs, harassment, threats, and sexual content.
Words, with a second opinion
Say it before
you send it.
A fast, context-aware profanity check that keeps up while you write.
Live check
Your text is checked as you type.
Signal
Ready when you are
Confidence
--
Type something above to see category scores
Jev evaluates context, not just keywords.
A tiny moderation stack
Small surface.
Thoughtful underneath.
Jevfanity turns one sentence into a set of focused moderation questions. That keeps the result legible: you get one overall signal, plus the category scores that explain it.
The API runs on Cloudflare Workers. The browser demo and the npm client both talk to the same small HTTP contract.
Read context
Long text is chunked, lookalike Unicode is normalized, and each category keeps its strongest signal.
Return detail
Applications get a simple flagged value, confidence scores, selected policy level, and usage metadata.
For fun, try a signal
Test the little
decision engine.
Take it with you