#11 jevmod content moderation
Community/app moderation: probability per category (spam, scam, harassment, nsfw, offtopic, self-harm, doxxing, minors, plus plain-language custom rules); you own thresholds and actions.
Fails open if Jev is unreachable (messages left alone). Self-harm is flag-only by design. Not affiliated with TypeSafe / Discord / Telegram / Reddit.
One yes/no (Noul) question per enabled category with explicit true/false criteria (pattern of TypeSafe guardrails cookbook; questions live in jevmod/categories.json)
spam- unsolicited promotion, invite farming, bare link drops, mass mentions
scam- fake giveaways, phishing domains, impersonated support, "DM me for a deal"
harassment- insults, slurs, threats, targeted abuse, in any language
nsfw- sexual or gore content for a general audience (below threshold means SFW)
offtopic- unrelated to channel_topic; needs a topic to mean anything
selfharm- the author is in crisis or considering self-harm; alert moderators, never punish
doxxing- reveals or hunts private data about a real person
minors- sexualises a minor or shows grooming behaviour
rule:<name>- your rule in plain language, exceptions included
FREE NITRO for the first 100!! claim at discord-gifts.ru/nitroaction flag, category scam ≈0.99; scores include spam≈0.98 (PyPI README)- Link: https://pypi.org/project/jevmod/
- Date: 2026-09-18
- Confidence: high
- Secondary: https://flaviocopes.com/jev/ (medium-low) · 2026-09-18
Community package by Omar Hernandez (MIT). Also: https://github.com/ohernandezdev/jevmod. Benchmark claims are package-authored — verify on your own traffic. Related paid starter: T2 Content moderation. Short note: Flavio sketches feed/content-filter style experiments (confidence medium-low) — distinct from jevmod; not a duplicate long card.