ByAIModerator | 1.20+ ,Java 21+ V6

New Systems Added

1. Discord Bot Integration

• The old webhook system has been completely removed.
• A full-featured JDA-based Discord bot has been added.
• All bot settings are managed via bot.yml.
• Enables two-way communication between Discord ↔ Minecraft.

1. Sending Minecraft Commands from Discord

• Commands like /mcban, /mckick, /mcmute, /mcunban can be used directly in Discord chat.
• Manage the server without being in-game.
• Permissions for which roles can access commands are configurable in bot.yml.

1. Interactive Punishment Review System

• When a punishment is issued, a notification is sent to the Discord staff channel.
• Notification includes Approve and Revert buttons.
• High-level staff can undo incorrect punishments with one click.
• Provides transparency and auditing.

1. Punishment Payment System (Atonement)

• For minor or first-time offenses, players can choose:
  • Wait out the punishment
  • Pay in-game currency via Vault to remove it
• Introduces a money sink to the server economy.
• Reduces staff workload for minor offenses.

1. "What Happened While You Were Away?" – Staff Report

• Staff with byaimoderator.feature.loginreport permission receive a summary message upon login.
• AI summarizes discussions, questions, and violations that occurred while offline.
• Staff gain instant awareness of server events.

1. Suspicious Command Detection (Exploit Protection)

• Players attempting many invalid commands in a short period receive a warning.
• This is not a punishment, only an alert.
• Detects malicious users before exploits can be abused.

1. Smart Feedback System (/aimod feedback)

• Staff can provide detailed feedback on AI’s incorrect decisions.
• Command collects:
  • What was the wrong decision?
  • What should have happened?
  • Why?
• Enables AI to learn faster and more accurately.

Bug Fixes

• Double Punishment Issue: Manual punishments were duplicated via Batch Moderation or /report. Resolved with ManualPunishmentManager.
• Blacklist & Whitelist Conflict: Words present in both lists now correctly prioritize whitelist.
• Missing prompts.yml Instructions: Added batch_moderation instructions to both tr.yml and en.yml.

Optimizations & Improvements

• Plugin size reduced from ~8.5MB → <1MB (~90% reduction).
• Required libraries now auto-download at server startup.
• Whitelist Enhancement: Added new words suitable for Turkish and game jargon; false positives significantly reduced.
• Bypass Permission: Users with byaimoderator.bypass or OP permission are exempt from all systems.

Hello everyone,

The newest version of ByAIModerator, V5.2.2, is now live! This update makes our AI moderation system smarter, more accurate, and more user-friendly. Here are all the details:

AI Improvements

• Slang, friendly jokes, and typos are now analyzed more accurately.
  
  
• False punishments have been minimized.
  
  
• Friendly banter and competitive in-game talk won’t be misunderstood anymore.

Advanced /aimod Help Menu

• All subcommands are now displayed in one help menu with detailed descriptions.
  
  
• Easier for new users, more organized for administrators.

Multi-Language Support (Lang Files)

• The /aimod dblist interface can now be managed via lang files (e.g., lang/tr.yml, lang/en.yml).
  
  
• Adapting to different languages is now much easier.

️ SuperiorSkyblock2 Compatibility

• Private island chat channels (/is chat) are now protected by ByAIModerator.
  
  
• This ensures all chat areas are covered by a single moderation system.

New Commands & Permissions

/aimod feedback

• Staff can instantly correct AI’s mistakes.
  
  
• The AI learns from feedback for permanent improvements.
  
  
• Permission: byaimoderator.command.feedback

/aimod dblist

• Manage the dynamically learned blacklist entries through a visual panel.
  
  
• Permission: byaimoderator.command.dynamicblacklist

/aimod draft

• AI-powered professional announcement creation.
  
  
• Preview + options: [Send] [Edit] [Regenerate] [Cancel].
  
  
• Stylish announcements are now just one command away.
  
  
• Permission: byaimoderator.command.draft

/aihelp

• Players can get instant AI-powered help about server rules, commands, or events.
  
  
• Uses your server’s predefined context for accurate answers.
  
  
• Permission: byaimoderator.command.aihelp

⚙️ Filtering & Whitelist Improvements

• The whitelist system has been fully fixed. It only affects the specified word, while the rest of the sentence continues to be analyzed.
  
  
• Innocent words will no longer get blocked.

AFK Bot Blocking System

• Suspicious AFK behaviors (block breaking, fishing, idle) are detected by the AI.
  
  
• Players receive a GUI verification challenge → must select the correct item.
  
  
• Wrong choice = instant kick.
  
  
• Correct choice = optional spawn teleport or kick.

Automatic Chat Cleaning

• Inappropriate messages are automatically removed after punishments.
  
  
• Periodic cleaning keeps the chat tidy and safe.

Timestamped Archiving

• Chat history is saved in the database with timestamps.
  
  
• Weekly archiving and database reset ensures performance is maintained.

⚙️ Performance & Stability

• AI connection is now stabilized even under heavy load.
  
  
• Various bugs fixed, moderation works more reliably.
  
  
• Runs smoothly without extra server performance impact.

Important Notes

• Make sure to configure prompts.yml, blacklist.yml, rules.yml, and lang files according to your server.
  
  
• config.yml and severities.yml remain intact, no reset required.

❤️ Thank You ByAIModerator V5.2.2 takes moderation to the next level. Thank you for your support!

ByAIModerator V5.1 Update

Hello server owners and forum members,
ByAIModerator version V5.1 has been released. This version brings both new features and improvements in cost/performance.

New Features

/filter add|remove|list (byaimoderator.command.filter)

Players can block words they don’t want to see.
For example, with /filter add lan, player Ahmet filters the word “lan”. When Burak types “naber lan”, the message will appear as “naber ***” for Ahmet.
The system can be enabled/disabled via config.yml, and the censor character can be customized.

Dynamic Blacklist
A message previously flagged as inappropriate by the AI will be blocked directly if repeated, without being resent to the AI.
This significantly reduces API costs.

Ignored Words & Min Length
With the ignored-words list, certain short words won’t be sent to the AI (e.g., ok, gg, tmm).
With min-length, you can set the minimum character length required for a message to be analyzed.

Punishment & Warning Messages

public-broadcast: Now, player-specific punishment messages can be disabled.

send-warnings-to-player: You can decide whether warning messages should be sent to the player.

Performance & Cost

The system now operates with lower costs.
The planned “send chat to AI once every 60 seconds in batch” feature was not included in this version due to issues in testing. However, this feature will be added in version V5.2 once it works flawlessly.