Compliance & Acceptable Use Policy

Last updated: 11 November 2026

This Compliance & Acceptable Use Policy ("Policy") applies to every customer of CYBERSHEILD PTE. LTD., a company incorporated in the Republic of Singapore, doing business as MIST Ai ("MIST Ai") and to anyone who uses a sub-account, API key or Output obtained through MIST Ai. It forms part of our Terms of Service. Capitalised terms have the meanings given there.

MIST Ai is an authorized reseller of Seedance on BytePlus ModelArk and Wan on Alibaba Cloud Model Studio, as permitted under its agreements with BytePlus and Alibaba Cloud.

1. Upstream acceptable use policies flow down to you

Your use of each model is also governed by the Upstream Provider's rules, as updated from time to time. You must comply with them as if you held the upstream account yourself. They include:

Seedance on BytePlus ModelArk

• BytePlus Acceptable Use Policy, including its section on AI services: https://docs.byteplus.com/en/docs/legal/docs-acceptable-use-policy
• BytePlus GenAI Acceptable Use Policy: https://docs.byteplus.com/en/docs/legal/acceptable_use_policy_byteplus_genai
• BytePlus Customer Agreement and Service Specific Terms: https://docs.byteplus.com/en/docs/legal/docs-customer-agreement, https://docs.byteplus.com/en/docs/legal/docs-service-specific-terms
• ModelArk content filter: https://ai.byteplus.com/ark/region:ap-southeast-1/docs/Content_Pre-filter

Wan on Alibaba Cloud Model Studio

• Alibaba Cloud International Website Product Terms of Service, section 4.48 (Model Studio): https://www.alibabacloud.com/help/en/legal/latest/alibaba-cloud-international-website-product-terms-of-service-v-3-8-0
• Alibaba Cloud International Website Membership Agreement (content restrictions): https://www.alibabacloud.com/help/en/legal/latest/membership-agreement
• Alibaba Cloud International Website Acceptable Use Policy: https://www.alibabacloud.com/help/en/legal/latest/alibaba-cloud-international-website-acceptable-use-policy

If this Policy and an upstream policy differ, the stricter rule applies. If an upstream policy changes, the new rule applies from the date the upstream change takes effect.

2. Prohibited content and uses

You must not use the Services, or any Output, to create, request, store or distribute any of the following, or to try to:

2.1 Sexual content

• (a) Any sexually explicit, pornographic or adult content, including nudity for sexual purposes, erotic role-play and "NSFW" generation. This applies even where the content would be lawful in your jurisdiction and even with age-gating. Adult-content platforms and services are not eligible customers.
• (b) Child sexual abuse material (CSAM), or any sexualised depiction of a minor or of a person who appears to be a minor, whether real, fictional, animated or AI-generated. We report suspected CSAM to the competent authorities and preserve evidence as required by law.
• (c) Non-consensual intimate imagery of any real person.

2.2 Real people, likeness and deepfakes

• (d) Depicting a real, identifiable person's face, body or voice without that person's documented consent, or the consent of their legal guardian or estate where applicable, or other lawful right.
• (e) Impersonating any person or organisation, or creating content that falsely appears to be real footage of real events, people or officials. This includes political deepfakes, fake endorsements and fake news footage.
• (f) Using Outputs for fraud, scams, phishing, identity verification bypass or financial deception.
• (g) Evading upstream face and likeness protections. For example, Seedance 2.x on BytePlus ModelArk rejects reference images and videos containing real human faces. Use only the upstream provider's documented compliant routes, such as authorised portrait-asset libraries or preset virtual characters. Never use workarounds such as cropping, masking or image laundering to get a real face past the filter.

2.3 Harm to people and society

• (h) Content that promotes or depicts terrorism or violent extremism; incites violence; or gives instructions for weapons, explosives or other dangerous activities.
• (i) Harassment, bullying, threats, hate speech or discrimination based on protected characteristics.
• (j) Glorification or encouragement of self-harm or suicide.
• (k) Human trafficking or exploitation; harm to minors in any form.
• (l) Disinformation intended to mislead the public, including about elections, public health or emergencies.

2.4 Rights and law

• (m) Content that infringes copyright, trademark, rights of publicity or privacy, or other rights. This includes uploading reference material you do not have the rights to use.
• (n) Processing sensitive personal data, such as health, biometric, financial or children's data, without a lawful basis.
• (o) Any activity that is illegal where you operate or where the Output is distributed.

2.5 Platform integrity

• (p) Bypassing, probing or disabling content filters, rate limits, safety systems or access controls, including jailbreak prompts aimed at prohibited content.
• (q) Reverse engineering the models, or using the Services or their Outputs to develop, train, distil or improve competing models or services.
• (r) Removing, hiding or tampering with watermarks, provenance metadata or AI-generated labels that upstream services apply.
• (s) Reselling, sharing or sublicensing sub-account access, or using our trade names or upstream trademarks to suggest endorsement.
• (t) High-risk uses where failure could cause death, injury or serious harm, including military, weapons and critical-infrastructure operations.

Mature fiction. Fictional violence, horror or other mature but non-sexual content may be submitted only if it is lawful in the places you operate and publish, and only to the extent it is not blocked by the Upstream Provider's filters. You must not attempt to circumvent those filters. Graphic violence or gore may still be blocked. Adult and sexually explicit content remains prohibited in all cases under Section 2.1.

3. AI-generated content: labelling guidance

Many jurisdictions now require, or are moving toward requiring, disclosure of AI-generated or synthetic media. You are responsible for meeting the rules that apply where you publish. At a minimum we expect you to:

1. Keep upstream labels. Do not remove watermarks, AI labels or provenance metadata, such as C2PA content credentials, added by the Upstream Services.
2. Disclose realistic synthetic media. Where Output shows realistic people, places or events, clearly tell viewers it is AI-generated or AI-altered, for example with an on-screen label, caption or description.
3. Follow platform rules. Use the AI-content disclosure tools of the platforms you publish on, such as YouTube, TikTok, Meta and X, where available or required.
4. Watch for specific laws where they apply. These include, where applicable, Article 50 of Regulation (EU) 2024/1689 (the EU AI Act), the Administrative Measures for the Labelling of AI-Generated Synthetic Content of the People's Republic of China, and other laws that require disclosure of AI-generated or synthetic media, including (where they apply to you) laws on election deepfakes and digital replicas. Singapore official guidance should be taken into account where relevant, but is not treated here as equivalent mandatory labelling law.
5. Tell your end users. If you build an app on the Services, tell your users that they are interacting with an AI system and that its content is AI-generated.

This guidance is not legal advice. Get your own advice for your markets.

4. Responsible deployment

If you integrate the Services into your own product or workflow, you must:

• test before going live and keep appropriate human oversight;
• use reasonable input and output moderation for your use case;
• give your users a way to report harmful content;
• apply age-appropriate access controls where minors could reach your product; and
• keep records needed to investigate abuse, for example request IDs and timestamps, consistent with your own privacy obligations.

5. Export control and sanctions

5.1 You must comply with all applicable export control and economic sanctions laws, including those of the United Nations and Singapore, and those that apply to you or to the use of the Upstream Services (which may include the United States, the European Union and the United Kingdom).

5.2 You represent that neither you nor your owners, directors, or end users who receive the Services:

• (a) are on any sanctions or restricted-party list;
• (b) are owned 50% or more, or controlled, by a listed party; or
• (c) are located in, organised under the laws of, or ordinarily resident in a comprehensively sanctioned country or region under those laws.

5.3 You must not use the Services for, or provide Outputs to, any prohibited end use, such as weapons of mass destruction or military end uses where restricted, or prohibited end user.

5.4 We may refuse, suspend or terminate service without liability where we reasonably believe sanctions or export rules apply. Upstream Providers may also restrict access by region.

5.5 The Services are offered only to customers outside mainland China, and only in the territory permitted under our authorization and the Upstream Provider's then-current availability rules.

6. Enforcement

If we believe this Policy or an upstream policy has been breached, we may, at our discretion and without liability:

• warn you or ask you to remove content;
• revoke API keys or restrict permissions;
• suspend or terminate your sub-account and Order Form;
• report to law enforcement or regulators, and keep and disclose information where the law requires; and
• cooperate with Upstream Providers' investigations.

Upstream Providers may take their own action, including blocking content or suspending our account. That can affect your access without notice from us.

7. Reporting abuse

To report content or activity you believe breaches this Policy, including deepfakes of you, IP infringement or suspected CSAM, contact:

Email: [email protected]

Please include, where you can: a description, URLs or file identifiers, timestamps or request IDs, and your contact details. For IP complaints, also include the work you own, the infringing material, and a good-faith statement that you are authorised to act.

We aim to acknowledge reports within 2 business days and to prioritise reports involving minors or imminent harm. We may forward reports to the relevant customer and/or Upstream Provider. Do not send CSAM itself to us. Describe where it is located, and report it directly to the police or to a recognised hotline in your country (for example, in the United States, the NCMEC CyberTipline).

Security vulnerabilities: [email protected].

8. Changes

We may update this Policy to reflect changes in law or in upstream policies. Material changes will be posted here with a new "Last updated" date.

9. Trademark notice

"Seedance", "BytePlus", "ModelArk", "Wan", "Alibaba Cloud" and "Model Studio" are trademarks or trade names of their respective owners. MIST Ai uses these names only to identify the services it resells under its agreements with their owners. No ownership of these marks is claimed, and customers receive no right to use them.