Acceptable Use Policy
Last updated: August 26, 2026
1. Scope
This policy applies to prompts, reference images, other inputs, and images, videos, audio, and other generated results submitted to or returned by Vistella. Vistella is operated independently by an individual and provides the publicly listed AI models and pay-as-you-go creative services described on this site.
2. Prohibited Content and Uses
You may not use the Service to generate, upload, distribute, or help create: - pornography, explicit sexual content, or NSFW content; - violence, gore, graphic violence, self-harm, or content encouraging harm; - hateful content or attacks against protected groups; - child sexual abuse material (CSAM) or any other unsafe content involving children; - deepfakes, non-consensual face swaps, impersonation of real people, or content intended to mislead viewers about its source; - content that infringes copyright, trademarks, publicity, privacy, or other third-party rights. You must not upload third-party material without permission, bypass prompt, image, or account safety controls, or claim that AI-generated content was created by a human.
2.1 Absolute Prohibitions and Zero Tolerance
The following are prohibited without an exception for creative, educational, research, or news context: - child sexual abuse material (CSAM) or any sexualized content involving minors; - promoting, recruiting for, assisting, or glorifying terrorism, extremism, or mass violence; - technical instructions for creating or improving weapons of mass destruction (WMD); - incitement to genocide, ethnic hatred, or violence against a group; - non-consensual sexual deepfakes, impersonation, or identity fraud targeting a real person. When detected, we may reject the request, restrict or terminate the account, preserve necessary evidence, and report to authorities when legally required or necessary to protect safety. Do not attach CSAM to a report.
2.2 Other Restricted Conduct
We may also refuse or restrict scams, phishing, malware, cyberattacks, unlawful weapons, false identities, harassment, defamatory abuse, spam at scale, and jailbreak prompts intended to evade safety controls. The Service currently provides no adult-content exception or age-verification channel. Requests involving politicians, celebrities, or brands must not mislead viewers about source, manipulate elections, or infringe rights.
3. AI Models and Disclosure
Examples listed on the current public models page include GPT Image 2, Nano Banana 2, Qwen Image 2.0, FLUX 1.1 Pro, Ideogram V4, and Seedream 5.0 Pro. Actual availability, providers, and parameters are shown on the model page and in the price estimate before generation. Outputs are produced by third-party models and may be inaccurate, duplicated, or biased; they are not factual statements or human-authored works by Vistella.
4. Automated Review, Human Review, and Logs
We screen every prompt and input parameter before generation, and safety-review generated images before they are displayed or returned. In production, an automated prompt safety check runs before queueing, billing, or model invocation; deny, flag, timeout, and service errors fail closed. If a public user gallery is enabled in the future, published content will also receive asynchronous post-publication review. Prompts also pass through a local safety word and rules list; reference images and generated images use an automated vision safety check. The safety list is actively maintained as risks and laws change, rather than being a one-time static list. Images are sent only temporarily during the review request and moderation metadata is retained for 90 days by default. If review fails or the result is uncertain, the content remains pending and is not directly published; flagged requests enter a human-review queue. The current Inspiration gallery is a curated static gallery maintained by Vistella, not a user publishing channel.
5. Enforcement
We may refuse or delay generation, remove content, limit or zero related credits, require deletion of material, suspend or terminate an account, and provide necessary information when legally required. Zero-tolerance categories receive priority action; we may preserve evidence and report when required. We retain necessary moderation logs, model decisions, and human-review outcomes to handle abuse, disputes, and safety incidents. You must not evade review through spelling variants, encoding, bulk requests, or other methods.
6. Reports and Security Contact
To report suspected illegal content, child-safety content, non-consensual deepfakes, copyright infringement, or a security vulnerability, email support@vistella.xyz with “Safety report” or “Security issue” in the subject. Include a task ID, page link, account email, or screenshot where possible; do not resend child sexual abuse material in the email. We normally acknowledge a report within two business days, prioritize child-safety, imminent-harm, and account-takeover reports, and aim to provide an outcome within seven business days. Complex matters or matters involving external authorities may take longer; we will explain the status.
7. Appeals and Review
If a request was incorrectly blocked or an account was restricted, email support@vistella.xyz with the account email, task ID, relevant time, and reason for the appeal. We normally acknowledge an appeal within two business days and aim to complete human review within seven business days; we will explain the expected timing for complex cases. We will review available moderation logs and may restore content or access when appropriate. An appeal is not guaranteed to succeed and does not suspend legal or platform safety requirements.
8. Policy Updates and Transparency
We actively update the safety list and this policy as laws, model-provider rules, abuse trends, and review findings change. Material changes are reflected by the date on this page and may be announced through the website or service notices. We do not disclose internal rules in a way that would help users evade review.