Moonoo.ai has implemented an internal outbound content policy filter that operates before content is transmitted to external AI or publishing providers.
The filter reviews user-provided prompts together with relevant contextual information, including the selected brand profile, products and services, campaign information, marketing strategy, existing asset content, and any text edited immediately before publishing.
Requests are blocked when they match prohibited or high-risk categories, including:
- Sexually explicit, pornographic, NSFW, adult-service, or sexual-gratification content
- Spyware, surveillance malware, keyloggers, telecommunications jamming, or manipulation equipment
- Illegal products, services, forged documents, fake identification, or fake-reference services
- Illegal drugs, synthetic drug alternatives, drug-production equipment, cannabis businesses, cannabis cultivation, and restricted THC products
- Illegal weapons, firearm accessories, explosives, radioactive materials, toxic materials, and other dangerous goods
- Content that promotes unlawful violence, physical harm, or violence against protected groups
- Unsafe pseudo-pharmaceutical or nutraceutical products making harmful or misleading medical claims
- Pyramid schemes, multilevel marketing recruitment schemes, get-rich-quick offers, guaranteed-return claims, fake testimonials, high-pressure upselling, and other deceptive or abusive practices
The filter is enforced at multiple points in the workflow:
- Before website content is used for AI-powered brand analysis
- Before initial AI content generation
- Before regenerating or modifying an existing asset
- Before generated content is sent to social networks, advertising platforms, email providers, or connected WordPress websites
When a request is blocked, the content is not sent to the external provider. The system displays a clear notice explaining that the requested content conflicts with our Terms. You may revise the request and try again or open a support ticket if you believe the decision was made in error.
Support tickets created from the moderation notice are automatically identified as content-policy reviews, providing an escalation path for manual review. We also record limited enforcement metadata, such as the matched policy category, user, and workspace, without storing the full blocked content in application logs.
These controls are implemented independently of any safety measures provided by the underlying AI model providers and are designed to be updated as prohibited-product and regulatory requirements evolve.