Levelable

A moderation system that uses trust levels and smart restrictions to reduce spam, limit abuse, and automate subreddit moderation.

Comment

Inspiration

Reddit moderation today is heavily reactive. Most communities rely on removals, bans, or manual approvals after low-quality or spam content has already been submitted. We were inspired by the idea that moderation should be progressive instead of binary.

We wanted to build a system where new users can still participate, but gradually unlock more privileges as they contribute positively to the community. The goal was to reduce spam and moderator workload without hurting onboarding for legitimate users.

What it does

Levelable is a Devvit-based progressive moderation system for Reddit communities.

The app introduces subreddit-specific trust levels that dynamically control what users are allowed to do. Depending on their trust level, users may have restrictions on:

  • posting frequency
  • comment frequency
  • image uploads
  • video uploads

As users participate positively, they earn trust and unlock more privileges automatically.

Moderators can configure:

  • trust thresholds
  • posting limits
  • media restrictions
  • probation states
  • cooldown durations

The system automatically monitors posts and comments, applies moderation rules in real time, removes restricted content, and explains moderation decisions transparently to users.

How we built it

We built Levelable using:

  • Devvit Web
  • TypeScript
  • Reddit Developer Platform
  • Devvit Redis
  • Hono for API routing

The system architecture is modular and built around several core services:

  • trust engine
  • rate limiter
  • media detector
  • enforcement engine
  • logging service
  • moderator dashboard

Redis is used to store:

  • trust profiles
  • rate limit windows
  • moderation logs
  • subreddit configuration

When users submit posts or comments, Devvit triggers evaluate:

  • trust level
  • posting frequency
  • comment frequency
  • media type
  • probation state

The enforcement engine then decides whether to allow or remove the content.

Challenges we ran into

One of the biggest challenges was designing a moderation system that balances spam prevention with user onboarding. We did not want the app to feel overly restrictive for legitimate new users.

Another challenge was designing subreddit-specific trust logic that remains flexible enough to support different community types, such as discussion forums, creative communities, and marketplace subreddits.

We also spent significant time thinking about fail-open behavior to ensure that moderation errors would never accidentally block legitimate content.

Accomplishments that we're proud of

We are proud that Levelable feels like a complete moderation framework rather than just a simple moderation bot.

Some accomplishments we are especially proud of include:

  • progressive trust-based moderation
  • configurable media restrictions
  • probation and cooldown states
  • subreddit-specific trust profiles
  • transparent user-facing moderation explanations
  • a scalable modular architecture

We are also proud that the project feels realistic and production-oriented while still being achievable within a hackathon scope.

What we learned

We learned that moderation is not just about detecting bad behavior — it is also about guiding good behavior.

We also learned how important transparency is in moderation systems. Clear explanations and progressive restrictions create a much better user experience than silent removals or immediate bans.

On the technical side, we learned how to design scalable event-driven moderation workflows using Devvit and Redis.

What's next for Levelable

In the future, we want to expand Levelable into a full community-state moderation platform.

Planned features include:

  • AI-assisted risk scoring
  • adaptive rate limiting during spam waves
  • reputation-based content prioritization
  • moderator analytics dashboards
  • automated strike decay
  • subreddit onboarding flows
  • cross-community reputation systems
  • smarter moderation recommendations

Our long-term vision is to transform moderation from a reactive cleanup process into a proactive community management system for Reddit communities.

Built With

  • devvit
  • hono
Share this project:

Updates