ModPulse AI is a Devvit-powered Reddit moderation triage tool that helps subreddit moderators analyze risky posts, detect scam/spam/harassment/privacy signals, generate risk scores, and produce structured moderation reports with suggested actions.
It is built for moderators who want faster, clearer, and more consistent review decisions directly inside Reddit’s moderation workflow.
Moderators often spend a lot of time reviewing posts that may contain:
- Spam or self-promotion
- Crypto or financial scams
- Misleading claims
- Off-platform solicitation
- Harassment or toxic language
- Privacy or sensitive information risks
Manual moderation can become repetitive and inconsistent, especially when communities grow and queues become larger.
ModPulse AI adds a custom moderator action to Reddit posts.
Moderators can open ModPulse AI Analyzer, enter the post title/body, and instantly generate a structured moderation report containing:
- Risk score
- Risk level
- Priority
- Recommended decision
- Detected categories
- Detected signals
- Evidence summary
- Suggested moderator actions
- User-facing removal message
- Estimated moderation time saved
- Devvit-based Reddit mod tool
- Post-level moderation action
- AI-ready moderation analysis
- Rule-based fallback scoring engine
- Risk score from 0 to 100
- Critical / High / Medium / Low classification
- Evidence-based moderation report
- Suggested moderator action checklist
- Ready-to-send removal message
- Moderator time-saving estimate
- Report modal inside Reddit
ModPulse AI can detect signals such as:
- Crypto or financial promotion
- Guaranteed profit claims
- Referral or discount link promotion
- Urgency-based spam
- DM or off-platform solicitation
- Harassment or abusive language
- Possible private information exposure
Risk Score: 95/100 Risk Level: Critical Priority: Immediate Recommended Decision: Remove
Detected Categories:
- Scam / Financial Risk
- Misleading Claim
- Spam / Self Promotion
- Off-platform Solicitation
Detected Signals:
- Financial or crypto promotion detected
- Unrealistic guarantee claim detected
- Referral or promotional language detected
- Urgency or off-platform contact detected
Suggested Moderator Actions:
- Remove the post immediately
Devvit TypeScript Hono Node.js Reddit Developer Platform OpenAI-ready architecture Rule-based fallback engine
Moderator opens a Reddit post. Moderator clicks Open ModPulse AI Analyzer from the post moderation menu. Moderator enters or reviews the post title and body. ModPulse AI analyzes the content. A structured moderation report is displayed. Moderator uses the risk score, evidence, and suggested actions to make a faster decision.
ModPulse AI is designed with OpenAI API support for AI-based moderation analysis.
If an OpenAI API key is not available, the app automatically uses a stable rule-based fallback engine.
Example PowerShell command:
$env:OPENAI_API_KEY="your_api_key_here" npm run dev
If OpenAI is active, the report shows:
Analysis Source: OpenAI
If fallback is active, the report shows:
Analysis Source: Rule-based fallback Run Locally
npm install
Build the project:
npm run build
Start Devvit playtest:
npm run dev
src/ index.ts # Main Devvit server and moderation analyzer logic
devvit.json # Devvit app configuration package.json # Project dependencies and scripts Commands npm run dev
npm run build
npm run deploy
npm run launch
Publishes the app for review and public use.
npm run login
Authenticates the Devvit CLI with Reddit.
ModPulse AI is designed for the New Mod Tool category.
Reducing moderator workload Improving moderation consistency Speeding up risky post triage Helping moderators identify scams, spam, harassment, and privacy risks Generating clear evidence-based reports Making moderation decisions easier for mod teams Impact
For high-risk posts, ModPulse AI can save an estimated 5–7 minutes per review by automatically preparing:
Risk analysis Evidence summary Action checklist Removal message
This helps moderators act faster and maintain better community safety.
Future Improvements Automatic post content fetching Full custom dashboard view Exportable PDF moderation reports Community-specific rule configuration Moderator analytics dashboard Auto-escalation for repeated offenders Integration with Reddit mod queue workflows Configurable subreddit-specific risk rules Author
Built by Mohamed Farhun M.
BSD-3-Clause
- Mark as spam if promotional or scam-related
- Review the author history for repeated violations
- Consider a temporary ban if this is repeated behavior