jevdit

Jev is an AI classifier made by TypeSafe. It checks every chosen username, subjevdit, post, and comment before publication. Automatic usernames come from curated word lists. Each subjevdit’s description tells Jev what belongs there. There are no human subjevdit moderators.

Clear rules. Clear feedback.

Our discussion guidelines favor curiosity, good faith, and substantive disagreement.

Be curious and civil. No harassment, hate, threats, sexual or exploitative content, private information, scams, malware, or spam. Subjevdit rules can be more specific, but they cannot override these ground rules.

If a submission is blocked, you’ll see its category and an explanation. Your draft stays in the form so you can revise it. If Jev is uncertain or unavailable, publication waits for a successful check.

Links get a closer look.

Links in posts, comments, and replies are checked. We fetch page titles, descriptions, and readable text so Jev can judge the content in context. For YouTube, we check the video title, channel, and available description—not the video itself. Linked content can change later; a check is not a guarantee.

Private by default.

You automatically start with an anonymous account when browsing. Click your username to link an email and keep access, or use sign in to return to an existing account. Email addresses are used for sign-in links and are never displayed publicly. Your public submissions and linked page content are sent to TypeSafe for moderation. Do not include secrets or personal information. Moderation records contain a content hash and classification, not a copy of rejected text.

A thoughtful experiment.

AI moderation can make mistakes. The checks and conservative thresholds reduce risk; they do not promise a network free of harmful content. You can revise a blocked submission and try again.