Acceptable Use Policy

Last updated: August 28, 2026

1. Scope

This Acceptable Use Policy ("Policy") applies to all use of NarraTwin, including reference images, article text, instructions, prompts, AI analysis, generated or edited images, stored projects, exports, reports, and other content processed through the Service. It forms part of our Terms of Service.

NarraTwin is an independently branded article-illustration product. It currently uses Google Gemini 2.5 Flash for article, visual, and output-safety analysis, and Google Nano Banana 2 for image generation and editing. These models are accessed through WaveSpeedAI. NarraTwin is not affiliated with, endorsed by, or an official product of Google, Gemini, Nano Banana, or WaveSpeedAI.

Generated images are AI-generated. You must review them before publishing or relying on them.

2. Prohibited generated content

NarraTwin prohibits requests to create, edit, store, or distribute each of the following categories through the Service:

1. Pornography or NSFW content — explicit sexual acts, sexualized nudity, fetish content, non-consensual intimate imagery, or sexual exploitation.
2. Violence or gore — graphic injury, mutilation, torture, cruelty, glorified extreme violence, or instructions that facilitate violent wrongdoing.
3. Hate speech — content that attacks, dehumanizes, threatens, or promotes hatred or extremist violence against people based on protected traits.
4. Child-unsafe content and CSAM — child sexual abuse material, sexualized depictions of minors, grooming, exploitation, or any content that endangers a child. This category is subject to zero tolerance.
5. Deepfakes or impersonation — deceptive, harmful, or non-consensual use of a real person's identity, face, voice, likeness, or alleged statements, including fraudulent evidence and non-consensual sexual deepfakes.
6. Copyright or trademark infringement — unauthorized reproduction or confusing misuse of protected characters, artworks, logos, brands, trade dress, or other intellectual property.

The following are also prohibited:

• terrorism, violent extremism, genocide advocacy, or weapons-of-mass-destruction instructions;
• fraud, fake identity documents, scams, spam, malware, or evasion of safety controls;
• unlawful surveillance, privacy or publicity-rights violations, harassment, exploitation, or disclosure of another person's confidential information;
• misleading political or public-interest content presented as authentic evidence; and
• any content or conduct prohibited by applicable law or a model provider's binding terms.

Creative, documentary, newsworthy, medical, or educational framing does not make CSAM, sexual exploitation, non-consensual intimate imagery, or instructions for serious wrongdoing acceptable.

3. Content moderation

NarraTwin operates platform-side moderation in addition to safety controls used by model providers.

Before generation

Text sent to an image-generation workflow is checked by Waffo Prompt Sift using rules and semantic safety analysis. A request may proceed only when the enforced verdict is allow. A block, review, unavailable-service, or invalid verdict fails closed: the image model is not called, and eligible reserved credits are returned.

After generation, before display

Every generated image is classified by Google Gemini 2.5 Flash through WaveSpeedAI before NarraTwin writes it to the user's private storage or displays it. The classifier assesses the six prohibited categories above. Images marked block or review, and images that cannot be classified reliably, are not stored or shown. Eligible credits are returned.

NarraTwin does not operate a public gallery or public feed. Content that a user exports and publishes elsewhere is subject to that destination's rules and is outside NarraTwin's direct control.

Audit records and human review

We retain limited moderation records such as timestamps, prompt hashes, Waffo request IDs, verdicts, matched policy categories, safety-model task IDs, and enforcement outcomes. Prompt Sift does not return or retain prompt text in its redacted verdict. Retention and deletion are described in our Privacy Policy.

Reports and appeals are reviewed by a person. We may inspect the relevant task, inputs, output, account history, and moderation records only as reasonably needed to decide the case, protect users, and meet legal obligations.

4. Reporting violations

Report suspected policy violations, child-safety concerns, impersonation, or infringement to support@narratwin.com or through our contact page. This is a monitored reporting channel.

Include, where available:

• the NarraTwin task, project, or account identifier;
• a link, screenshot, or description of the content;
• the policy category involved;
• whether you are the affected person or rights holder; and
• any urgent safety or preservation concern.

We limit access to reports, protect reporter information where reasonably possible, and assess reports impartially. Knowingly false, abusive, or duplicative reports may themselves lead to restrictions.

5. Response targets

Response times depend on completeness, complexity, and legal process. Our operational targets are:

We may act immediately before contacting the affected user when delay could increase harm. Complex matters may take longer; when appropriate, we will tell the reporter that review remains open.

6. Enforcement

Depending on severity, history, intent, and legal requirements, NarraTwin may:

• block a prompt or generated image before display;
• warn the user or require changes;
• remove stored content or disable an export;
• restrict models, features, generation volume, or credits;
• reverse credits obtained through abuse;
• temporarily suspend or permanently terminate an account;
• preserve evidence, notify an affected provider or rights holder, or cooperate with a lawful investigation; and
• report apparent criminal conduct to an appropriate authority where required or reasonably necessary to protect people.

Confirmed CSAM, non-consensual sexual content involving real people, terrorism, or credible imminent harm may result in immediate permanent suspension and legal escalation. Attempts to bypass moderation are separate violations.

7. Appeals

You may appeal an automated block or account-level enforcement decision within 14 days by emailing support@narratwin.com. Include the task or account identifier, the decision being appealed, and relevant context. We target a response within 10 business days.

An appeal is reviewed again using the full available record and, where practicable, by a reviewer who did not make the original manual decision. An appeal does not require us to restore unlawful content or content that would expose users, providers, or NarraTwin to unacceptable risk.

8. Policy updates

We may update this Policy as the Service, models, safety systems, or law changes. The current version and update date will remain publicly available. Material changes will be communicated where reasonably practicable or legally required.

9. Contact

Safety reports, infringement notices, and appeals: support@narratwin.com or our contact page.