Morddit - Mod Triage Assistant

Morddit auto-analyzes every post and comment, routes them into HIGH/MEDIUM/LOW priority queues, and lets mods take one-click action - all without leaving Reddit.

Comment

Inspiration

Reddit mods are overwhelmed. Between spam, rule violations, doxxing, and hate speech, a busy subreddit can see hundreds of posts a day that need human review. Most mods rely on instinct and luck — there's no systematic way to prioritize what to look at first. We built Morddit to give every mod team an always-on triage assistant that does the first pass automatically.

What it does

Morddit auto-analyzes every post and comment the moment it's submitted to your subreddit. It scores content against your custom rules using keyword matching and 6 built-in red-flag heuristics (hate speech, doxxing patterns, spam links, excessive caps, and more). Each item is routed into a HIGH / MEDIUM / LOW priority queue — so mods always know where to look first.

Key features:

  • Always-on analysis — every post scored automatically, no manual steps
  • Priority queue — HIGH violations surface first; reported posts auto-escalate
  • Duplicate detection — Jaccard bigram similarity across the last 500 posts
  • One-click mod actions — Approve, Remove, Spam, Lock with pre-filled reasons
  • Plain-English rules — type rules in your own words; keywords extracted automatically
  • Works on install — 5 default rules seeded immediately, zero configuration needed
  • Auto-dismiss — when a mod acts, the item clears from the queue automatically

How we built it

Built entirely on Devvit (TypeScript) with Reddit-managed Redis for persistence. No external APIs, no backend server, no database costs. The analysis engine uses two complementary signals: keyword matching against mod-defined rules, and 6 hardcoded red-flag pattern checks. Scores map to HIGH (≥50), MEDIUM (≥20), or LOW (<20). Duplicate detection uses Jaccard similarity over 2-word bigrams against a 500-post rolling index. All UI lives inside a single Devvit custom post (the Dashboard), with four tabs: Overview, Rules, Queue, and Log.

Challenges

Getting reliable background processing working on Devvit required engineering around platform constraints — we implemented a scheduler-based polling approach that auto-analyzes new posts every 60 seconds, seeded directly from the Dashboard UI on first open.

What's next

Natural language rule suggestions, comment thread analysis, ban pattern detection across users, and multi-mod collaboration features.

Built With

Share this project:

Updates