Content Policy

Principles

Seedance 2.5 is built for legitimate commercial and creative work. Our automated safety layer is intentionally narrow: it blocks only high-confidence hard redlines, while normal creative ambiguity and mature atmosphere are allowed through to the selected model provider.

We do not review text while you type. Content checks happen when you submit a generation request or upload a reference image.

Content We Do Not Support

You may not use Seedance 2.5 to upload, generate, request, publish, or distribute:

• Child sexual abuse material, sexual content involving minors, or sexualized depictions of minors.
• Explicit adult sexual content, including non-consensual intimate content.
• Realistic graphic violence, mutilation, torture, or cruelty.
• Terrorist or extremist recruitment, praise, instruction, or propaganda.
• Sensitive secrets such as private keys, API keys, access tokens, or credentials.
• Malware, fraud, spam, impersonation, or content intended to harm people, systems, or accounts.
• Content that violates intellectual property, privacy, publicity, or other legal rights.

Creative Context

We allow ordinary creative work, including action scenes, fictional conflict, stylized combat, historical or documentary topics, fashion, swimwear, sensual editorial styling, romantic atmosphere, implied intimacy, satire, and artistic expression, as long as it does not cross the hard redlines above.

Mature or suggestive direction is allowed when the subjects are clearly adult and consensual. NSFW is not a blanket ban: if a prompt or reference is blocked, adjust it and generate again. If a draft lands in a sensual or semi-suggestive style, that alone is not treated as abuse or a penalty event; the hard line is explicit sexual content, minors, coercion, or other unsupported content.

Political, religious, historical, and social topics are not automatically blocked. Deceptive use of real public figures, including misleading deepfakes or false endorsements, is handled through policy review, reports, and manual enforcement rather than broad automated political filtering.

Prompts and Reference Uploads

Prompts are checked for high-confidence hard-redline categories before credits are consumed or a generation task is created. If a prompt is blocked, no credits are charged.

Reference image uploads may be checked before storage. Video and audio references may also be reviewed by upstream model providers, and we may add frame-based video checks where needed for safety and compliance.

Enforcement

A single blocked request is not a punishment. You can rewrite the prompt or remove the unsupported reference and try again.

Repeated hard-redline attempts may trigger temporary generation cooldowns or admin review. Zero-tolerance categories involving minors may be escalated for urgent review and legal compliance.

Appeals and Reports

If you believe a request was incorrectly blocked, contact us through the support ticket center at /settings/tickets while signed in, or use the contact page for abuse and content review concerns.

When reporting content, include the account email, task ID or URL when available, the reason for the report, and any relevant context.

Relationship to Provider Rules

Seedance 2.5 also relies on upstream model and infrastructure providers. Even when our first safety gate allows a request, a provider may reject it under its own rules. Failed generation tasks are handled by the existing task and credit refund flow when the system can confirm the failure.