Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.aurinfer.com/llms.txt

Use this file to discover all available pages before exploring further.

​
AI Moderation

The Moderation tab provides access to a highly sophisticated, AI-driven server protection system. Rather than relying on simple keyword blacklists, AurInfer reads the context, tone, and intent of messages to decide if they violate community rules, automatically taking action based on your exact specifications.

​
1. Moderation Configuration

Control how strictly the AI guards your community and what rules it should enforce.

​
Moderation Mode

Determine the AI’s overall tolerance level for borderline behavior:
  • Lenient: The bot only steps in for severe, unambiguous violations (e.g., direct threats, massive spam).
  • Moderate (Default): The balanced approach. Enforces rules reasonably while allowing sarcasm or friendly banter.
  • Strict: The “zero tolerance” setting. The AI acts on minor rule breaks, bad language, or slight hostility.

​
Defining Server Rules

You must tell the AI what rules to enforce. You have two options:
  1. Rule Text Area: Write out your rules manually directly in the dashboard UI.
  2. Fetch from #rules: Automatically import and sync the rules by pointing the AI to your server’s dedicated rules channel.

​
Per-Rule Actions

Once rules are listed, you can map specific punishments to each individual rule. The AI can execute any of the following actions autonomously when a rule is broken:
  • WARN: Sends a direct warning to the user.
  • TIMEOUT: Temporarily mutes the user.
  • KICK: Removes the user from the server (they can rejoin).
  • BAN: Permanently bans the user.
Example configuration:
  • Rule 1 (Be Kind) -> Mapped Action: TIMEOUT
  • Rule 2 (No Spam) -> Mapped Action: WARN
  • Rule 3 (No NSFW) -> Mapped Action: BAN

​
Moderation Instructions

This instruction box defines the personality and limits of the “Moderator AI”. You can use this to say things like:
  • “You are a fair judge, do not ban anyone without notifying the admin channel first.”
  • “Ignore mean comments if they are directed exclusively at you, you are a tough AI.”
  • “Warn users publicly in the channel where the rule was broken, publicly embarrassing them.”

​
Advanced Moderation Settings

Below the instructions, you have fine-grained control over exemptions, automated action scaling, and logging:
  • Personal Bias: Toggle whether the AI character develops its own natural bias/opinions about users based on historical interactions.
  • Warn Before Action: Ensures the AI always issues a warning message before taking any stronger disciplinary steps.
  • Exemptions:
    • Excluded Users: Specific Discord User IDs that the AI will unconditionally ignore during moderation.
    • Excluded Roles: Allow entire Discord server roles to bypass AI moderation (e.g., admin or trusted roles).

​
Automated Moderation Actions

Toggle the master switches for what the AI is legally allowed to do. If any of these are disabled, the AI will never perform that specific action, even if a rule calls for it.
  • WARN: Allows the AI to send warning notices.
  • TIMEOUT: Allows the AI to timeout users.
    • Warning Count: Set a threshold (e.g., 3 warnings) required before a timeout triggers automatically.
    • Duration: Length of the timeout in minutes (e.g., 120 mins).
  • KICK: Allows the AI to kick users from the server.
  • BAN: Allows the AI to issue permanent server bans.

​
Notifications & Mod Logs

Keep your human staff in the loop when the AI takes action.
  • Warning Delivery Method: Choose between Channel Warnings (public shaming) or Direct Messages (private).
  • Mod Channel: Select a dedicated channel (e.g., #mod-logs) where the AI stores permanent records of its moderation actions.
  • Notify Server Admins/Mods: When a suspicious situation occurs, the AI can ping your team.
    • Alert Channel: Select the destination channel for alerts.
    • Role to Ping: Choose a specific staff role to mention when an alert triggers.

​
2. Audit Logs

The Audit Logs act as a searchable database of every single action the Moderator AI has ever taken, ensuring complete transparency. For every disciplinary action, the audit log captures:
  • Action Type: (Warn, Timeout, Kick, Ban)
  • User ID: Who was punished.
  • Reason: Which rule was broken.
  • Message Content: The exact text the user sent that triggered the action.
  • AI Explanation: A detailed reasoning paragraph outlining why the AI believed the message broke the specific rule.
  • Timestamp: Exactly when this occurred.
This log is invaluable for overturning false positives and fine-tuning your rules based on AI interpretation.

​
3. Warnings Management

The Warnings interface is a dedicated management page specifically for active warnings. Here, your human moderator team can:
  • View all active warnings stacked against users.
  • Read the context of the warning.
  • Revoke warnings. If the human moderation team disagrees with the AI’s ruling, you can easily clear the warning from the user’s record with a single click.