AI Use Policy

This AI Use Policy applies to all VideoDescriber AI-assisted features, including video analysis, video-to-prompt generation, subtitle and audio processing, content description, and video generation. It supplements our Terms of Service and Privacy Policy.

Your Rights and Responsibilities

You may upload, link to, analyze, transform, or generate only content that you own or are legally authorized to use. You are responsible for obtaining all necessary permissions, licenses, and consents for submitted videos, images, audio, voices, faces, text, prompts, and other materials.

You are also responsible for reviewing AI-generated outputs before publishing, distributing, or relying on them. VideoDescriber does not verify ownership before every task, grant rights in third-party content, or guarantee that an output is accurate, original, non-infringing, authentic, or suitable as evidence or professional advice.

Content Standards and Prohibited Categories

You must not use VideoDescriber to create, upload, analyze, transform, distribute, or facilitate:

• Pornographic, adult, or NSFW content — Any nudity, sexually explicit or sexually suggestive content, sexual acts, fetish content, sexual services, or other material intended for sexual arousal. This prohibition applies whether the content is real, synthetic, consensual, artistic, or animated.
• Child sexual exploitation or endangerment — Child sexual abuse material, sexualized depictions of minors, grooming, or any other exploitation or sexualization of children. Apparent or AI-generated minors are treated as minors under this policy.
• Non-consensual intimate content — Intimate or sexual imagery involving a real or identifiable person without their explicit authorization.
• Graphic violence and serious harm — Graphic gore, torture, mutilation, glorification of extreme violence or cruelty, unlawful threats, harassment, stalking, or instructions intended to cause serious harm.
• Hate and discrimination — Content that attacks, dehumanizes, promotes discrimination against, or incites violence toward protected people or groups.
• Deepfakes, impersonation, and fraud — Phishing, scams, or deceptive synthetic media that misrepresents a person, organization, event, product, endorsement, or evidence.
• Intellectual-property and other rights violations — Copyright, trademark, privacy, publicity, confidentiality, or other rights violations, including using content without required authorization.
• Deceptive documents and records — Falsified identification, certificates, receipts, invoices, financial records, legal evidence, safety information, or other verification materials.
• Illegal or harmful technical activity — Malware, unauthorized surveillance or access, evasion of safeguards, or any other illegal activity.

You must not remove or conceal watermarks, attribution, copyright notices, provenance information, content credentials, or authenticity indicators without authorization. The technical ability to process content does not give you permission to use it.

Synthetic Media and Disclosures

Do not use generated or altered media to deceive people about its origin or authenticity. You must obtain consent when using a recognizable person's face, voice, likeness, or personal information and disclose synthetic or materially altered content whenever required by law, platform rules, or the circumstances in which it is presented.

Safety Measures and Review

Before a video-generation request is sent to an AI model, VideoDescriber uses Waffo's content-safety service to scan the final prompt in enforcement mode. Generation proceeds only when Waffo returns an allow decision; prompts returned for review, blocked prompts, and service-degraded results are not sent to the generation model. As described in the Waffo Prompt Scan documentation, Waffo provides a redacted decision and audit request ID and states that its scan is stateless, meaning the raw prompt is not retained after the response. We retain the decision metadata with an approved generation task for compliance and support purposes.

VideoDescriber and its AI providers may also use automated safeguards, rate limits, provider-level safety controls, and review of reported or flagged activity. These systems may not identify every violation and do not transfer responsibility for your content or use to us.

When reasonably necessary to operate or protect the service, investigate abuse, enforce this policy, protect users or third parties, or comply with law, we may review relevant prompts, uploads, outputs, task metadata, and account activity in accordance with our Privacy Policy.

Enforcement

Depending on the nature and severity of a violation, we may refuse or stop processing, remove content, restrict features or credits, suspend or terminate accounts, preserve relevant records, and cooperate with lawful requests. We may act without prior notice when reasonably necessary to protect users, children, rights holders, third parties, our providers, or the service.

Current AI Models and Providers

Video generation currently uses ByteDance Seedance 2.5, Kling AI Kling 3, and Google Veo 3.1, accessed through Fal. Depending on the quality selected, VideoDescriber may use Kling 3 Standard or Pro and Veo 3.1 Fast or its standard endpoint. The exact model selected for a generation task is shown in the video generator before submission.

Other features may use different models or local browser processing. Providers, model versions, and endpoints may change as service availability, quality, and safety requirements evolve; when the available video-generation models change, we will update the generator and this policy. AI outputs can be incomplete, inaccurate, or unexpected, so you must independently evaluate them before use.

Reporting Violations

Our monitored reporting channel for content or conduct that may violate this policy is [email protected]. Use the subject line “AUP Report” and, where available, include:

• A concise description of the suspected violation and the relevant prohibited-content category.
• The related VideoDescriber task ID, public URL, account identifier, and approximate date and time.
• Screenshots or other supporting context that can be shared lawfully. Do not email, reproduce, or attach illegal material, including child sexual abuse material; instead, describe where it appeared.
• An email address where we may request clarification. We will limit disclosure of a reporter's identity to what is reasonably necessary to investigate, protect affected people, and comply with law.

After receiving a report, we will acknowledge receipt when contact information is provided, triage the report based on severity, and preserve or temporarily restrict relevant content or account activity when reasonably necessary. We may then use automated checks and human review to assess the available task metadata, prompts, inputs, outputs, account activity, and supporting evidence in accordance with our Privacy Policy. Confirmed violations may result in the enforcement measures described above, and critical matters may be referred to the relevant service provider or lawful authority.

We prioritize reports involving child safety, credible threats, non-consensual intimate content, or imminent harm. We aim to begin reviewing other sufficiently detailed reports within two business days. Where legally and operationally appropriate, we will notify the reporter when the review is closed or request additional information; complex investigations may require more time.

Contact

Questions about this policy can be sent to [email protected].