QueueLens

QueueLens gives Reddit moderators a native review card with policy signals, author context, and optional AI analysis. The moderator makes the call.

Comment

Inspiration

Moderating a borderline post is rarely just reading the post. You end up checking the author, karma, reports, edits, history, local rules, and the signals that tell you whether something is useful, promotional, risky, or unclear.

That context work matters, but it is scattered.

QueueLens was built to organize the first look. The goal is simple: context without control. Give moderators useful review context without taking the decision away from them.

What it does

QueueLens is a first-look review assistant for Reddit moderators.

From the Mod Shield post menu, a moderator can run QueueLens on the current post and get a compact review card. The card shows a recommendation, confidence, reason, author context, post context, policy signals, optional AI advisory, notes, and feedback.

The local analyzer runs first and works without AI. QueueLens AI is optional and advisory only.

QueueLens does not approve, remove, ban, report, comment, or enforce rules. The moderator decides every time.

How we built it

QueueLens is built as a Reddit-native Devvit app using Mod Shield post actions, Devvit server endpoints, subreddit settings, Redis caching, native forms, and a local signal analyzer.

The analyzer checks post and author context against the active subreddit policy profile. Redis caches review cards for speed, with refresh and invalidation paths for policy changes, post edits, and manual refresh.

The optional AI path is bring-your-own-key. Moderators provide their own Gemini key. If Gemini fails, the local card still appears.

Challenges we ran into

The hardest parts were Mod Shield routing, cache safety, and keeping the optional AI layer from becoming a dependency.

A stale review card is worse than no card, so cache invalidation needed care. The AI layer also had to fail cleanly every time.

Signal wording was another challenge. QueueLens has to be useful without pretending certainty.

Accomplishments that we're proud of

QueueLens works inside Reddit’s native moderator workflow. No external dashboard is required.

The standard review path works without AI. The card is advisory only. The refresh path lets moderators regenerate a current card. Policy profiles let different communities review posts by their own rules.

Most importantly, QueueLens stays bounded. It helps a moderator make a better decision without becoming the decision-maker.

What we learned

Moderators do not want automation replacing judgment. They want leverage.

The useful product is not the biggest AI system. It is the smallest trustworthy surface that helps a moderator make a better decision.

What's next for QueueLens

Next is signal tuning, expanded policy profiles, multi-post author context, and better review reporting for mod teams.

QueueLens supplies context. The moderator decides.

Built With

Share this project:

Updates