Overview

Rabu AI uses automated content filtering combined with human review to prevent harmful content from being generated or shared on our platform. This page explains what categories of content are permanently blocked, how our filtering works, and what to do if you believe content was blocked in error.

Permanently Blocked Content Categories

The following content is always blocked on Rabu AI. These blocks are absolute — they cannot be unlocked by any subscription tier, account setting, or special request:

CSAM and content involving minors — Any sexual or sexualized content involving real or fictional minors. This includes lolicon, shotacon, or any attempt to describe AI characters as under 18 in sexual contexts.
Non-consensual depictions of real people — Sexual or intimate content depicting real, identifiable individuals without their explicit consent.
Bestiality — Any sexual content involving real or realistic animals.
Content promoting violence against real people — Graphic depictions of violence, torture, or harm directed at specific real individuals or groups of real people.
Terrorism and extremism — Content that promotes, glorifies, recruits for, or facilitates acts of terrorism, genocide, or violent extremism.
Doxxing — Sharing, requesting, or generating content that reveals private personal information (home address, phone numbers, workplace details, etc.) about real individuals without consent.
Illegal drug production instructions — Detailed instructions for manufacturing controlled substances, illegal drugs, or dangerous chemicals.
Content designed to harass specific real individuals — Content clearly intended to target, intimidate, humiliate, or harass a specific real person.
Weapons manufacturing instructions — Detailed instructions for producing illegal weapons, explosives, or other devices designed to cause mass harm.
Human trafficking facilitation — Any content that facilitates, promotes, or enables human trafficking or forced exploitation.

How Our AI Content Filters Work

Rabu AI uses a multi-layered content filtering system:

Input filtering — User messages are analyzed before being processed to detect requests for prohibited content.
Output filtering — AI-generated responses are reviewed before being displayed to catch any prohibited content that slipped through input filters.
Pattern detection — Our systems recognize common prompt injection techniques, jailbreak attempts, and indirect ways of requesting blocked content.
Contextual analysis — Filters evaluate the broader context of a conversation, not just individual messages, to catch multi-step attempts to circumvent restrictions.
Continuous improvement — Our filter systems are regularly updated based on new threats, user reports, and evolving content norms.

Human Review

Automated filters handle the vast majority of content decisions. However, human reviewers from our trust and safety team review:

Escalated user reports
Edge cases where automated systems flag uncertainty
Appeals from users who believe their content was blocked in error
Patterns that suggest coordinated abuse attempts

False Positive Appeals

Our filters are designed to err on the side of caution, which means legitimate content is occasionally blocked. If you believe your content was blocked in error, you can appeal:

Email [email protected] with subject line "Content Appeal"
Describe what you were trying to do and why you believe it should be allowed
Include your account username

We review appeals within 3–5 business days. Note that appeals for content in the permanently blocked categories listed above will not be approved — those restrictions are absolute.

Reporting Content That Should Be Blocked

If you encounter content on Rabu AI that you believe violates our policies and should be blocked, please report it immediately:

Email [email protected] with subject line "Content Violation Report"
Include a description of the content and, if possible, screenshots
Reports involving minors or illegal content are treated with highest priority

Consequences for Attempting to Bypass Filters

Attempting to circumvent our content filters — including through prompt injection, roleplay framing, character manipulation, or any other technique — is a violation of our Terms of Service and may result in:

Immediate account suspension
Permanent ban for serious or repeat attempts
Reporting to law enforcement for attempts to generate illegal content

Blocked Content Policy