Jev in production › Moderation and content filtering

X Reply Judge (Jev)

Chrome extension uses Jev to score the probability each X post or reply is AI-written, badging it red, orange or gray by threshold.

Open on GitHub ↗

The source publishes no measured number.
Use
Moderation and content filtering
Industry
Media and social
Form
Open-source tool
Stage
Announced
Plugs into
X
Listed
2026-10-20
Found via
github
Repository
liaocaoxuezhe/x-reply-judge-jev
Stars
2
Forks
1
Last push
2026-10-21
Language
JavaScript
License
none stated

The README opens with

Chrome MV3 extension that adds one small badge to each post or reply on X (Twitter): AI 87%, an experimental estimate that the text was AI-written, as judged by Cloudflare Clef / Clef Flash or TypeSafe Jev. Colors: = 70 red, = 31 orange, otherwise gray. No reasons, no signal lists.

Also adds Mute / Block buttons on each reply (post detail / notifications). They act only when you click them.

This project is a modified derivative of dontbesilent2025/x-reply-guard (commit 56bda1b), Copyright (c) 2026 dontbesilent, MIT License. The DOM scraping (data-testid selectors), cache/auto-scan flow and the hide/block logic in reply-cleaner.js come from that project. See LICENSE.

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for moderation and content filtering