Acceptable Use Policy

2. Content Standards and Prohibited Categories

You may not generate, upload, store, publish or distribute the following six categories through the Service:

• Sexual / NSFW content.
• Violence / gore.
• Hate speech.
• Child sexual abuse material (CSAM).
• Deepfakes / impersonation of real people.
• Content infringing copyright or trademarks.

2.1 Absolutely Prohibited (Zero Tolerance)

The following content is actioned immediately and may be reported to the competent authorities:

• Child sexual abuse material (CSAM), or any sexualised content involving minors.
• Content promoting or facilitating terrorism, extremism or mass violence.
• Technical instructions for manufacturing weapons of mass destruction.
• Content inciting genocide or ethnic hatred.
• Deepfakes of real people, including face swaps, nudity generation, intimate scene synthesis, and fabricated statements or behaviour — whether or not the person consented.
• Other content that violates the laws and regulations of the People's Republic of China or mandatory rules in your jurisdiction.

2.2 Restricted Content (Requires Review)

The following is permitted only under specific conditions; we reserve the right to review case by case and refuse:

• Violent content: must have a clear creative, educational or documentary context and must not glorify or incite real-world violence.
• Politically sensitive content: must not contain misleading statements, disinformation or election interference.
• Medical, legal or financial advice: must not impersonate a professional providing licensed advice, and must comply with local regulations.

2.3 Generally Prohibited Conduct

• Generating or spreading disinformation, debunked claims, or fabricated news, official documents or evidence.
• Impersonating a real person, brand or organisation, or implying authorisation or endorsement.
• Infringing intellectual property, likeness, privacy, reputation or other lawful rights.
• Presenting AI-generated content as human-created in order to deceive others.
• Bulk-generating spam, fraudulent marketing content or fake reviews.
• Using jailbreak prompts or other means to bypass platform safety mechanisms.
• Using the Service's inputs, outputs or model responses to train, fine-tune, distil, benchmark or otherwise develop a competing model.

2.4 Source Requirements for Uploaded Assets

You may upload only your own photos, videos or voice, or material for which you have explicit authorisation from the person depicted. Uploading another person's photo or voice without consent (including public figures) is prohibited.

Before uploading, you must confirm that the source is lawful and that the scope of authorisation covers the generation and use you intend; authorisation does not override the prohibition on deepfake content in Section 2.1.

Uploading material depicting minors requires explicit consent from their guardian and must never be used in improper or sexualised scenarios.

3. Content Moderation

We operate a multi-layer moderation system: automated review, human review and periodic audits, with audit logs retained for verification.

3.1 Automated Review (Real Time)

• Input filtering: keyword, semantic and intent detection on prompts; Waffo Prompt Sift performs a safety review before generation, and requests matching prohibited categories are blocked with a reason shown to the user.
• Output control: image and video outputs rely on the upstream model providers' built-in safety filters, combined with sampling and report-driven review.
• Usage analysis: monitoring for anomalous patterns such as bulk registrations, automated volume abuse and retry behaviour that attempts to evade moderation.
• Safety models: Waffo Prompt Sift (prompts) and upstream providers' built-in safety classifiers (outputs).

3.2 Human Review

• Requests automatically flagged for review enter a human review queue and no result is returned until review completes.
• Reported content is reviewed manually by the content safety team.
• Reviewers receive content safety training and sign confidentiality agreements.
• For high-risk requests we may require clarification, restrict features or refuse to generate.

3.3 Periodic Audits

• Regular random sampling of generated content.
• Continuous evaluation and improvement of classifier and sensitive-word accuracy.
• Logging automated results, human review conclusions and enforcement actions for compliance audits and appeals.

3.4 Severity Levels

4. Reporting

If you find content or behaviour that violates this AUP, report it through the channels below:

4.1 What to Report

• Content or behaviour violating the content standards in this AUP.
• Serious violations involving child safety or terrorism.
• AI impersonation of you, or infringement of your intellectual property, likeness or privacy.
• Other urgent situations requiring platform intervention.

4.2 Reporting Principles

• Confidentiality: reporter identity is kept strictly confidential.
• Neutrality: every report is reviewed independently and objectively, regardless of who is reported.
• Anti-abuse: malicious or false reports are recorded and handled according to severity.

5. Reporting Channels

Report email: support@draftio.ai (the same monitored mailbox as support and appeals; please put "Content Report" in the subject line).

In-app reporting: use the "Report" entry on a work detail or share page.

We acknowledge reports automatically and complete triage and handling within the timeframes in Section 6.

6. Response and Handling Timeframes

If handling requires more time, we proactively tell the reporter why and give an estimate. Process: receive (auto-acknowledge) → triage (L1–L4) → review (automated + human) → decide → enforce → notify the parties.

7. Enforcement Actions

We may also deduct or zero out compute points and benefits obtained through violations and cancel related orders, except where mandatory law requires a refund.

8. Appeals

If you believe an enforcement action was wrong, email support@draftio.ai within 15 days of the notice, with material supporting your position.

We complete the review and inform you of the outcome within 5 business days; the review is conducted independently by staff who were not involved in the original decision. The original action generally remains in effect during the appeal.

9. Transparency and Updates

• We assess and, where appropriate, publish content safety transparency information annually.
• Material changes are announced at least 15 days in advance via an in-app announcement or registered email, and the "Last updated" date at the top of the page is revised.
• Continued use after the effective date constitutes acceptance of the updated AUP.

10. Contact Us

• Safety and reporting: support@draftio.ai
• Appeals: support@draftio.ai
• Customer support: support@draftio.ai
• Company: Shenzhen Xinghui Future Technology Co., Ltd.
• Registered address: Room 501, Building A, Ganfeng Technology Building, Bantian Street, Longgang District, Shenzhen, China
• Website: https://draftio.ai

Contact us

If you have questions about this page, account security, data processing, appeals or complaints, contact us at support@draftio.ai. We reply within 2 business days.

Shenzhen Xinghui Future Technology Co., Ltd. · Room 501, Building A, Ganfeng Technology Building, Bantian Street, Longgang District, Shenzhen, China · https://draftio.ai