Acceptable Use Policy

Effective date: October 20, 2026

About This Policy

SceneOak is an AI video generation service provided by COD Co., Ltd. (株式会社ＣＯＤ) ("we", "us"). This Acceptable Use Policy explains what you may not create or do with SceneOak, how we moderate content, how to report a violation and what we do when this policy is broken. It is part of our Terms of Service.

Every video generated with SceneOak is AI-generated content, not real footage.

Scope

This policy applies to everything you submit to or create with the service — prompts, chat messages, uploaded images and other media, and the videos generated for you — and to any other use of the service, including sharing a conversation through a public link. It applies whether or not the content is ever published.

Your clips are rendered by third-party AI model providers. Their usage policies also apply, and they may refuse content that this policy does not mention.

Prohibited Content

You may not use the service to create, upload, request or share any of the following.

1. Sexual and NSFW content — nudity, sexually explicit acts, pornography, and sexualised or fetish depictions of any person, whether real or fictional.
2. Violence and gore — graphic violence, gore, mutilation or realistic depictions of serious injury or death; content that promotes or glorifies self-harm, suicide or eating disorders; and content that praises, promotes or supports terrorism, violent extremism or mass violence.
3. Hate speech and harassment — content that attacks, demeans or dehumanises people based on race, ethnicity, national origin, religion, caste, sex, gender identity, sexual orientation, disability, age or serious illness; hateful symbols used to promote hate; and content that harasses, bullies, threatens or intimidates anyone.
4. Child sexual abuse material and child safety — any content that sexualises minors, including child sexual abuse material (CSAM), sexualised images of real or fictional minors, and content made to groom or exploit children. We have zero tolerance for this category: see Enforcement below.
5. Deepfakes and impersonation — animating a photo of yourself or of someone who has given you permission (for example natural movement, facial expressions or camera motion) is allowed. You may not depict a real, identifiable person without their consent. Whatever the purpose and even with consent, you may not: create any sexual or nude depiction of a real person (including "undressing", face swaps into intimate scenes and other intimate imagery); swap a real person's face onto someone else's body; make an identifiable real person, especially a public figure, appear to say or do something they did not say or do in a way presented as real; or impersonate another person, brand or organisation.
6. Copyright and trademark infringement — content that infringes someone else's copyright, trademark or other intellectual property rights, including uploading images, footage or artwork you do not have the right to use, and using another party's logos or brand to mislead people about who made or endorses the content.

Images You Upload

Only upload images you took yourself or have the right to use. If an image shows a real person, you must have that person's permission to animate it (or, for a minor, their parent's or guardian's permission). Permission never makes prohibited content acceptable: you may not use an uploaded photo of anyone to create sexual or nude content, or any other content listed under Prohibited Content.

Other Prohibited Uses

You also agree not to:

• Use the service for any illegal activity or to break applicable laws or regulations.
• Use the service for fraud, scams, spam, malware or other deceptive or harmful purposes.
• Attempt to gain unauthorised access to our systems or data, or interfere with or disrupt the service.
• Try to get around safety filters or model refusals, for example by rewording a refused prompt to obtain the same prohibited result.
• Use the service in breach of the usage policies of the AI model providers that render your clips.
• Pass off a generated video as real footage or as human-made work in order to mislead people.
• Use the service to create or spread disinformation.

How We Moderate

Content is moderated in four layers.

• 1. Pre-generation check. Before a prompt, image or video is sent to a video model, and before the credits for the video are charged, we check the prompt, every image the render would use (opening and closing frames and reference images) and three frames of every reference video (at about 10%, 50% and 90% of its length), using OpenAI's moderation model (omni-moderation). In text it detects sexual content, sexual content involving minors, violence and graphic violence, self-harm, hate, harassment and illicit activity; in images and video frames it detects sexual content, violence and graphic violence, and self-harm. If anything is flagged, the request is refused, nothing is generated and no credits are charged for the video, and we keep a record of the refusal (the account, the conversation and the categories flagged). If the check cannot run — for example because a reference video's frames cannot be read or the frame service is unavailable — the video is not generated either. Reference videos must be smaller than 99 MB, and are most reliably checked as MP4 files encoded with H.264. Images and videos must be uploaded to SceneOak first — links to files on other sites are not accepted — so what we check is exactly what the model receives. Reference audio is not covered by this check.
• 2. Check of the finished video. Before a finished video is stored or shown to you, three of its frames (at about 10%, 50% and 90% of its length) are checked by the same moderation model for the image categories above. If a frame is flagged, the video is not delivered, the credits for it are returned to your balance, and we keep a record of the refusal. If this check cannot run, the video is not shown until it can: it is checked again automatically while the conversation is open. If it still cannot be checked two hours after the render started, the next time the conversation is opened its delivery is cancelled and its credits are returned.
• 3. Model provider filters. Prompts and files that pass our check are also subject to the automated safety filters of the model that renders the clip — the third-party models listed in our Terms of Service, such as Kling, MiniMax and Seedance. These filters may refuse a prompt or file. When a render is refused or fails, it is not delivered, and the credits for that render are returned to your balance automatically. See our Refund Policy.
• 4. Review of reports. Our team reviews every report we receive and acts on it as described under Enforcement. To do so, we may examine the reported account's prompts, chat messages, uploaded images and videos, and generated videos.
• Action on our own initiative. We may also act without a report when we become aware of a likely violation, for example through a refusal recorded by our checks, a model refusal, a support conversation or a payment provider.
• Purpose. We review content only to enforce our Terms of Service and this policy and to prevent abuse, as described in our Privacy Policy. Within our team, only the staff who need to see your content to handle a review access it.

How the checks process your content. Video frames are taken by Cloudflare, which hosts SceneOak, and are not stored. The prompt, images and frames are sent to OpenAI, which processes them only to classify them; OpenAI states that it does not use requests to its moderation endpoint for training or retain them for abuse monitoring.

Automated checks can make mistakes in both directions — they may block acceptable content or miss prohibited content. They look only at the frames described above, so content that appears only between them, and any audio, is not checked automatically. You remain responsible for what you create and share.

Reporting Violations

Anyone can report content or behaviour that breaks this policy — you do not need an account.

• Email: services@bizcod.co.jp. Please start the subject line with "Abuse report".
• Support ticket: if you have an account, open a support ticket after signing in.

To help us act quickly, please include:

• the link to the content (for example, the shared conversation link or the video's address), or a clear description of it and where you saw it;
• which part of this policy you believe it breaks, and why;
• your contact details, so we can follow up;
• for a copyright or trademark complaint: your name, the work or mark concerned, a statement that you own it or are authorised to act for the owner, and a statement that the information in your report is accurate;
• for a deepfake or impersonation complaint about yourself: enough information for us to confirm that you are the person shown or their authorised representative.

Please do not send us copies of child sexual abuse material. Send the link or a description only.

We keep the identity of the person who reports confidential and do not share it with the reported user unless the law requires it. Knowingly false or abusive reports may themselves lead to action under this policy.

How We Handle Reports

1. We aim to acknowledge your report within the time shown in the table below.
2. We review the reported content and, where needed, the account behind it.
3. We decide whether the content breaks this policy and take the actions described under Enforcement.
4. We tell you that the report has been handled and whether action was taken. For privacy reasons we may not share details about the other user or the specific action.

If a report needs more time — for example because we need more information from you — we will let you know.

Enforcement

When we find a violation, we may take one or more of the following actions, depending on how serious it is, whether it was deliberate and whether it has happened before. We will tell you by email which action we took and why, unless the law or the safety of others prevents it.

• Refusal. Our pre-generation check or the model provider may refuse a generation, our check of the finished video may withhold it, and we may stop one.
• Warning. We may warn you and ask you to stop or to remove content you have shared.
• Content removal. We may take down a conversation. Any generation still running in it is stopped, it can no longer receive new messages, its public share link stops working, the videos generated in it and the images uploaded only to it are deleted from our storage, and it no longer appears in your history or library. We keep a record of the conversation for handling appeals and legal obligations.
• Suspension. We may temporarily suspend your account. While suspended you cannot sign in or generate videos, the conversations you shared publicly can no longer be opened from their links, your subscription is canceled so no further charges are made, and your credits are frozen and cannot be used.
• Termination. We may permanently terminate your account. As set out in our Terms of Service, we do not refund any payment, unused credits or the remaining period of your plan, and your remaining credits are forfeited.
• Reporting to authorities. We may report illegal content and the information needed to investigate it to law enforcement and other competent authorities, and we cooperate with lawful requests.

Zero tolerance. For the following, we remove the content and terminate the account immediately and without a warning, and we may report it to the competent authorities:

• child sexual abuse material or any content that sexualises minors;
• any sexual or nude depiction of a real person, including face swaps and "undressing", whether or not that person consented;
• content that promotes or assists terrorism, violent extremism or mass violence;
• content that incites genocide or ethnic hatred.

We report child sexual abuse material to the competent authorities, such as the police in Japan and, where relevant, the National Center for Missing & Exploited Children (NCMEC).

This does not limit rights you have under laws that cannot be excluded by contract.

Appeals

If you believe we made a mistake, reply to our notice email or email services@bizcod.co.jp within 30 days of the action, explaining why. We will review the decision and aim to reply within 10 business days of receiving your appeal. If we find we were wrong, we will reverse the action, for example by restoring your account and its remaining credits, which can be used until they expire. We do not reinstate accounts terminated for child sexual abuse material.

Changes to This Policy

We may update this policy. We will post the updated policy and its effective date on this page and, for material changes, notify you through the service or by email.

Contact

This policy is operated by COD Co., Ltd. (株式会社ＣＯＤ). Our registered address and other company details are listed in our Legal Notice (特定商取引法に基づく表記).

For reports, appeals or questions about this policy, email us at services@bizcod.co.jp or open a support ticket after signing in.