Inspiration

Most moderation tools are built to watch content or users. But who is watching the moderators? Behind every thriving subreddit is a human team that gets tired, takes vacations, or sleeps. When a team goes quiet, the community suffers. We wanted to build a tool that treats the subreddit like a living patient—monitoring its "heartbeat" (human actions) and ensuring that no community is ever left unattended, even when the humans are away.

What it does

Dr. Mod is a comprehensive "Subreddit Health Dashboard" powered by 12 specialized "Powers" in a centralized Control Room:

  1. 🩻 Sub Diagnosis (The Hero): A real-time 0–100 health score with a one-line diagnosis headline.
  2. 🏥 Crisis Detection: Monitors the "pulse" of human mod actions. If the team goes silent, the sub enters CRISIS.
  3. 🤖 AI Surgeon: Automatically takes over during a Crisis to approve or remove new posts based on your "Treatment Style."
  4. 🧑‍⚕️ Private Doc: A second pair of eyes that reviews present mod actions and suggests Auto-Rule improvements.
  5. 🩺 Mod Team Vitals: A real-time chart of who is active, idle, or flatlined.
  6. 📬 Mod Team Care: Automated, gentle modmail check-ins for teammates who haven't been seen in a week.
  7. 🔥 Burnout Watch: Predictive analytics that flag moderators who are at high risk of "flatlining" based on activity trends.
  8. 🌡️ Sub Temperature: A toxicity thermometer measuring the community's "fever" in real-time.
  9. 🎯 Find Good Moderators: An AI-powered talent scout that identifies the best potential mod candidates from your own community.
  10. 🗣️ User Appeal Flow: Author-to-mod appeal system for AI-automated removals.
  11. 📊 Weekly Health Report: A Sunday summary modmail of everything that happened.

How we built it

Our development journey was driven by one core problem: subreddit abandonment. We began with a simple Heartbeat Monitor to track moderator activity, but quickly hit a wall. Our initial "Featherless AI" minimalist approach—relying solely on basic heuristics—proved insufficient to handle the nuance of real-world moderation during downtime. We realized that to truly solve the problem of an unattended subreddit, the app needed a smarter cognitive layer.

This led to our first major pivot: integrating Google Gemini to act as an autonomous decision engine. This transformed the app from a simple stopwatch into a "Surgical AI" capable of maintaining community standards when humans are offline. Finally, we engineered a unified Dashboard Overlay inside the Live Monitor—not just for aesthetics, but a technical necessity to centralize the 12 distinct telemetry signals—from toxicity "Temperature" to "Burnout Watch"—into a single, actionable Control Room for the mod team.

Challenges we ran into

Integrating complex dashboard overlays inside the Reddit mobile webview was a challenge, specifically ensuring that static assets loaded reliably and that the "Heartbeat" animation performed smoothly at 60fps. We also spent significant time refining the "Diagnosis" logic to ensure it wasn't a "surveillance" tool, but a supportive "wellness" tool.

Accomplishments that we're proud of

We are incredibly proud of the Sub Diagnosis engine. It rolls dozens of signals (toxicity, pulse, burnout, team activity) into a single, intuitive number. We also achieved a "Zero-Touch" design—once configured, Dr. Mod can protect a subreddit entirely on its own if the mod team disappears.

What we learned

We learned that the most important part of a moderator's job isn't just removing bad content—it's maintaining a healthy, sustainable team. We learned how to balance AI automation with human oversight through our "Private Doc" supervision system.

What's next for Dr. Mod

We plan to expand the "Auto-Rule Doctor" to push suggestions directly to AutoModerator and add a "Triage Nurse" feature to automatically categorize and priority-sort the modmail inbox.

Note: Dr. Mod is the autonomous brand identity for the mod-doctor technical engine. A unified health substrate designed to ensure every subreddit under its care never misses a beat.

Built With

Share this project:

Updates