Acceptable Use Policy

This policy is part of the Terms of Service. It applies to everything you put into Pocky and everything you do with what comes out. Breaching it is a breach of the Terms, and is grounds for suspension or termination without refund.

Because Pocky runs on Anthropic's Claude models, your use is also subject to Anthropic's Usage Policy. Where that policy is stricter than this one, it applies. Nothing here permits something Anthropic prohibits.

Pocky writes drafts. It has no knowledge of your product, no access to your test results, and no ability to verify anything it says about what your product does. When you publish a brief, you become the advertiser and every claim in it is yours to stand behind.

You must not use Pocky to create, and must not publish:

• Claims about performance, results, ingredients, sourcing, certification, awards, pricing or availability that you cannot substantiate.
• Claims that a product diagnoses, treats, cures, mitigates or prevents any disease, or any other claim that would make a product a drug or medical device, unless you hold the approvals to make it.
• Weight-loss, income, earnings or investment-return claims that are not supported by evidence of typical results.
• Comparative claims about a competitor that you cannot support.
• Anything that is false, misleading, or deceptive by omission.

This is the single highest-risk use of a tool like Pocky, and it is absolutely prohibited. You must not use Pocky to generate, and must not publish or commission:

• A review, rating or testimonial attributed to a person who does not exist, or who never used the product.
• A review written as if by a customer, an employee, or a relative of either, without disclosing that connection.
• A testimonial that misrepresents what a real customer actually said or experienced.
• Content designed to suppress, bury or misrepresent honest negative reviews.
• An endorsement by a creator, expert, celebrity or organization that has not actually endorsed the product.

If you pay, gift, discount or otherwise compensate a creator, the material connection must be disclosed clearly and conspicuously in the content itself. Putting the disclosure in a brief is your job; Pocky will not add it for you.

Generating fake reviews and testimonials is separately unlawful in the United States and carries substantial civil penalties per review. We will terminate an account for it on the first instance.

Do not use Pocky to produce or promote:

Illegal and regulated

• Anything unlawful, or that promotes or facilitates unlawful activity.
• Weapons, explosives, illegal drugs, or controlled substances marketed outside their lawful channel.
• Tobacco, vaping, alcohol, gambling, cannabis, firearms, prescription drugs, supplements, financial products, or any other regulated category, unless you comply with every rule that applies to advertising it in your market and on the platform where it runs.
• Multi-level marketing recruitment, get-rich-quick schemes, or offers whose economics depend on recruiting rather than selling.
• Cryptocurrency or investment offerings that promise returns or omit required risk disclosure.

Harmful and abusive

• Content that is hateful, harassing, threatening, or that demeans a person or group on the basis of a protected characteristic.
• Sexual content, sexually suggestive content, or any content that sexualizes a minor. Content of this last kind is reported.
• Content that exploits or targets minors, including marketing directed at children.
• Content promoting self-harm, disordered eating, or dangerous behavior.
• Content designed to shame, frighten or manipulate a person into buying, including fabricated scarcity, fabricated countdowns and fabricated social proof.

Deceptive and infringing

• Impersonation of a person, brand or organization, including using a real person's name, voice, likeness or persona without permission.
• Content that infringes copyright, trademark, trade dress, or any other intellectual property right, including material copied from a competitor.
• Phishing, malware, scams, spam, or content intended to obtain credentials or payment under false pretenses.
• Political advertising, election content, or content designed to influence a vote or a public policy outcome.
• Content that discloses another person's private information without their consent.

Against the service

• Attempts to circumvent plan limits, quotas, cost controls or paywalls, including sharing an account across businesses or creating multiple accounts to obtain more free briefs.
• Scraping, automated access outside our published interfaces, or reselling access to the service.
• Using Output to train, fine-tune or evaluate a competing AI model or product.
• Probing, load testing, or attempting to bypass our security or our model provider's safety systems, including prompt injection aimed at producing prohibited content.
• Submitting other people's confidential information, or personal information you have no right to share, as brand material.

Pocky output is AI-generated. Some platforms, jurisdictions and advertising standards require you to disclose that content was created with AI assistance, and more will over time. Complying with those requirements is your responsibility as the publisher. If you are using Pocky output in a context where people could reasonably believe they are reading a human-written first-hand account, say that it is not.

We do not pre-screen what you generate. We act on reports, on safety refusals and signals surfaced by our model provider, and on patterns visible in usage data. Depending on severity we may warn you, remove content, restrict a feature, suspend the account, or terminate it. For the most serious categories, including content sexualizing minors and coordinated fake reviews, we terminate immediately and report where the law requires.

Termination for breach does not entitle you to a refund. See the Refund and Cancellation Policy.

If you believe someone is using Pocky in breach of this policy, or you have found content that appears to have been generated with it and should not have been, email ops@usepocky.com. Include links and enough detail for us to find it. We review every report we receive and we will not tell the account you reported who reported it.