Welcome to the Hangar Open Beta. Please report any issue you encounter on GitHub!
A PaperMC plugin that watches chat, batches messages, and asks an AI model to flag toxicity, spam, etc.
AiModeration 1.0.0
Initial release.
Highlights
- AI-powered chat moderation using OpenRouter or Groq
- Configurable model intelligence tiers (LIGHTWEIGHT, SMART, INTELLIGENT)
- Automatic batching: chat is analyzed every 20 messages
- Console-dispatched
warn/kick/banactions based on AI output - Permanent action audit log (
chat-action.log) - Optional Discord webhook embeds for every AI action
/aimoderationcommand withreload,status, andforcesend- Gson resolved dynamically via PaperMC's
librariesloader — no shading - Uses Java's built-in
HttpClient, no bundled HTTP dependencies
Requirements
- Java 21
- PaperMC 1.21.1 (API version 1.21)
Configuration
See GUIDE.md for full setup instructions.
Known Limitations
- Only OpenRouter and Groq are supported as providers.
- Punishment commands assume standard
kick/banconsole commands are available; custom punishment plugins are not yet supported. - No built-in rate limiting on API calls beyond the 20-line batching threshold.
Checksums
Generate your own checksum after building:
sha256sum target/AiModeration-1.0.0.jar
Platforms
Paper
1.21–26.2