Octavia's moderation capabilities rely on an advanced AI model which processes every message sent to your community, allowing us to analyse the message to determine whether it breaks your communities rules.

Setup

You can configure the moderation settings under the moderation tab in the dashboard. Select the appropriate action for each category, and Octavia will automatically moderate your chat.

Categories

To help you determine what categories you wish to moderate, below we outline the rules for categorisation.

Hate

Hate - High

• Slurs
• Threats against a person
• Incitement to Violence against any group of people.
• Credible Self-Harm Threats
• Support for Violent Extremism or Terrorism

Hate - Medium

• Incitement to violence against any person
• Covert Extremism
• Support for known hate group
• Hatred towards protected group
• Low Tier Slurs

Hate - Low

• General Racism, Sexism not otherwise covered
• Implied hate towards protected group
• Other low-tier hateful content.

Sexual

• Sexual Content not otherwise covered above

FUD, Shill and Pitch

Shill - Direct

• Outright Spam / Shilling for a project

Shill - Indirect

It is recommended to not moderate this category except when under attack.

• "Contextual" Spam
• Promoting other projects in a non-spammy way
• Other general soft shilling

FUD - Direct

• Outright claiming developers are lieing/cheating etc.
• Overt FUD around price/project success.

FUD - Indirect

It is recommended to not moderate this category except when under attack.

• Covert FUD
• Pretending to be part of the community / "expressing genuine concern" that is blatant

Pitch

• People pitching their services, or attempting to contact admins for purposes of selling services.

Other

Religion

• Content designed to start or maintain a religious conversation

Politics

• Content Designed to start or maintain a poltiical conversation (excluding crypto topics)

Price

• We are rolling out price moderation for communities unable to permit this discussion for regulatory reasons. It is being rolled out to all groups this February.

Non-Bannable Categories

We use these categories to provide analytical services to you and improve our services.

• Greeting
• General
• Crypto Conversation
• Community Conversation
• Project Conversation
• Support Conversation

Roadmap

In future we'll be improving our services further to include advanced analytics about your community over time. We're already collecting this data, so we're excited to expand our analytical services to help you grow your community. Additionally, we'll allow you to be more specific with your rules -- by adding text based rules to modify and allow/disallow different types of content to give you that granular control.

Usage

You can set the banned content in your dashboard, and the consequence for violations. We recommend muting for 5 minutes as a maximum for violations, as we are able to filter violations within 150ms in most cases.