Legal
Terms of Use
The agreement between you and ockeeper Labs for the studio, the vault and the API. Written to be read — not to be survived.
The short version
Your renders are yours
Commercial rights transfer to you the moment a render completes, and they survive cancellation.
We do not train on your work
Your prompts, uploads and outputs never enter a training set unless you opt in from settings.
Clear content rules
Six categories are off limits — sexual, violent, hateful, child-unsafe, deceptive and infringing content. Anyone can report a breach, and a human reviews every enforcement.
Accepting these terms
These Terms of Use are an agreement between you and ockeeper Labs, Unipessoal Lda., registered at Rua da Prata 80, 1100-420 Lisbon, Portugal. They apply to the studio at ockeeper.com, the vault, our mobile apps, the render API, and anything else we point back at this page.
By creating an account, generating an image, or sending a request to the API with your key, you accept these terms. If you are accepting on behalf of a company, you confirm you are allowed to bind that company — and “you” then means that company.
Your account
You need one account per person. Keep your password and API keys to yourself; anything done with your credentials is treated as done by you, until you tell us otherwise at security@ockeeper.com.
You must be at least 16, or the age of digital consent where you live, whichever is higher. Team workspaces can invite members, and the workspace owner is responsible for what those members render and for the credits they spend.
Credits, plans and billing
Rendering consumes credits. Subscription plans include an allowance granted at the start of each billing period, and one-time top-up packs add credits that do not expire while your account is open. The price and the credit cost are shown before you confirm, and any applicable tax is itemised at checkout based on your billing country.
Payments are processed by our payment provider, Waffo. Card details are entered on the provider’s hosted checkout page; ockeeper never receives or stores your full card number. We store only the order reference, the amount, the currency and the last four digits shown on your receipt.
Subscriptions renew automatically on the same day of each billing period until you cancel. Cancel at any time from billing settings: the next charge stops immediately, your plan keeps working until the end of the period you already paid for, and partial periods are not prorated.
If a render fails on our side, the credits return to your balance within an hour. If it fails because a safety filter refused the prompt, the credits are returned too. A render you simply dislike is still a render. For anything else, our refund policy is in section 4 and you can reach us at billing@ockeeper.com.
Refunds and cancellation
You may request a refund within 14 days of a charge by writing to billing@ockeeper.com with the order number. A purchase whose credits are unused is refunded in full. A purchase that is partly used is refunded in proportion to the credits still on the account.
Approved refunds are returned to the original payment method, normally within 5 to 10 business days depending on your bank. We confirm every refund by email, including the amount and the reference.
Where you are a consumer in the EU or UK, this policy is offered in addition to your statutory rights and does not limit them. Because credits are digital content delivered immediately, you agree that supply begins as soon as the purchase completes; any credits you have already spent are outside the statutory withdrawal right, and the balance is refundable as described above.
Content standards and prohibited categories
ockeeper generates images and video from the prompts and reference material you supply, and you keep full responsibility for what you type into a prompt and what you upload. The following six categories are prohibited across every surface of the service — the studio, the vault, the gallery and the API — whether the content is generated, uploaded as a reference, stored or published:
- Pornography and NSFW material — sexually explicit or pornographic imagery and video, including depictions of real or synthetic adults in explicit sexual acts, exposed genitalia, and sexualised nudity generated for arousal. Non-sexual nudity in an artistic, medical or educational context is assessed on its merits, but the service is not a platform for adult content.
- Violence and gore — realistic depictions of extreme violence, torture, mutilation, dismemberment, graphic injury or death; imagery glorifying or instructing self-harm or suicide; content promoting or celebrating terrorism or mass violence; and instructional depictions of weapons or explosives intended to cause harm.
- Hate speech — content that attacks, degrades, dehumanises or promotes discrimination or violence against a person or group on the basis of race, ethnicity, national origin, religion, caste, sex, gender identity, sexual orientation, disability or serious disease; hate symbols and the iconography of hate groups used to promote them.
- Child-unsafe content (CSAM) — any sexual or sexualised depiction of a minor, real or synthetic, including sexualised renders of a real child, of a fictional child character, or of an adult styled to appear as a minor. This is an absolute prohibition with no artistic, satirical or fictional exception, and it applies to prompts and reference uploads as well as outputs.
- Deepfakes and impersonation — synthetic likenesses of a real, identifiable person created to deceive: fabricated photorealistic depictions of real events, non-consensual intimate imagery of any person, fake endorsements or statements attributed to a real individual or organisation, and identity-locked Character Keys built from a private person without that person’s written consent, which we may ask you to produce.
- Copyright and trademark infringement — reproducing or closely imitating a work, character, style-plus-identity, brand, logo or trade dress you do not own or have no licence to use; uploading third-party copyrighted material as a reference without the right to do so; and generating counterfeit goods, packaging or brand assets.
Public figures may be depicted in clearly satirical or editorial contexts, but not in a way a reasonable viewer would read as a real photograph of a real event. Where a depiction of a real person is permitted and could plausibly be mistaken for a photograph, you must label it as AI-generated when you publish it.
These categories are not exhaustive. We prohibit anything unlawful where you or we operate, and anything that would put our payment providers or model partners in breach of their own rules. If you are unsure whether a project falls inside a category, ask legal@ockeeper.com before you render.
Reporting violations
Anyone — whether or not they hold an ockeeper account — may report content they believe breaches these standards. You do not need an account to file a report, and every route below reaches the same review queue:
- By email: write to abuse@ockeeper.com with the render or share link, or a screenshot, and a short description of the problem. Include a contact address if you would like to hear the outcome.
- Urgent child-safety reports: write to abuse@ockeeper.com with “CSAM” in the subject line. These bypass the standard queue and are triaged immediately, at any hour.
- Legal and rights-holder notices: write to legal@ockeeper.com for copyright, trademark or right-of-publicity claims, or where a court order or regulator is involved.
We acknowledge every report within 1 business day, and we tell the reporter the outcome once a decision is made, unless disclosure would compromise an investigation or a legal obligation. Reports may be submitted anonymously; we never share a reporter’s identity with the account being reported.
Rights holders reporting copyright or trademark infringement should send a notice to legal@ockeeper.com identifying the protected work, the infringing render, and a statement that you are the rights holder or authorised to act for them. We operate a repeat-infringer policy and terminate accounts that accumulate valid notices.
How we moderate content
Moderation runs at three points. Before a render, every prompt and reference upload is submitted to a third-party prompt-screening service and to the safety filters of the model provider you selected; a request matching a prohibited category is refused before any compute is spent, and no credits are charged for a refused prompt. After a render, outputs are scanned automatically, and material matching a prohibited category is blocked from the vault and the gallery. After publication, anything reported through the channels above, or surfaced by our own proactive scanning, enters human review.
Automated systems are not the last word. Every enforcement action that removes content or restricts an account is confirmed by a trained human reviewer before it takes effect, except where the material is confirmed child sexual abuse material, which is removed immediately and automatically. Reviewers see the render, the prompt, the model version and the account history.
- First or minor violation: the content is removed and you receive a written warning explaining which category it fell under.
- Repeat or serious violation: generation is suspended while we review, and the account may be limited to read-only vault access.
- Severe violation — CSAM, non-consensual intimate imagery, credible threats of violence, or a pattern of deceptive deepfakes: the account is terminated permanently, the content is removed from all systems, and unused credits are forfeited.
- Illegal material: we preserve the evidence required by law and report it to the competent authorities, including NCMEC and the relevant national body for child sexual abuse material.
We tell you what was removed and why, except where the law forbids notice. Every decision can be appealed to appeals@ockeeper.com within 30 days — a human who was not involved in the original decision reviews the appeal and replies within 1 business day. If an appeal succeeds, the content is restored and any credits consumed are returned.
Ownership of your renders
You own the images you generate, including the right to use them commercially, to modify them and to license them onward. That transfer happens when the render completes and is not conditional on your plan staying active.
We keep a licence to store your renders so the vault can serve them back to you, and to display a render publicly only if you explicitly publish it to the gallery. Nothing is used for marketing without a separate, written yes from you.
Model outputs are not unique. Another user with a similar prompt may receive a similar image, and neither of you can claim exclusivity over the other.
The vault and your data
Every render is written to your vault with the prompt, seed, model version and parameters that produced it. That history is what makes a render reproducible, so we keep it for as long as your account exists.
You can delete any item at any time; deletions clear from live systems immediately and from backups within 35 days. Export your full vault as a signed archive whenever you want, including after cancellation — read-only access to your history remains free. How we handle personal data is described in the Privacy Policy.
Acceptable use of the service
Beyond the content standards in section 5, there are things you may not do to the service itself:
- Resell raw model access, or wrap the API as a competing generation product without a written reseller agreement.
- Scrape the gallery, the vault or the docs with automated tooling outside the documented API.
- Circumvent credit limits by cycling free accounts, sharing keys across organisations, or spoofing region checks.
- Upload malware, run security testing without scope approval, or probe infrastructure you were not invited to test.
- Use the service for payment fraud, money laundering, or to buy credits with a card you are not authorised to use.
Models and third-party providers
Every generation model offered through ockeeper is operated by a third-party provider — OpenAI, Google, ByteDance, Black Forest Labs, Kuaishou, Alibaba and others — and each engine is labelled with its provider in the model picker. We do not train, fine-tune or host generation models of our own: your prompt is routed to the selected provider’s API, and that provider’s own usage policies apply on top of these terms. Where a provider imposes extra restrictions, those restrictions are shown in the model card before you render.
Models are versioned. We keep a version available for at least 180 days after it is superseded, and vault items always record the exact version that produced them so a re-render is honest about drift.
API access and rate limits
API keys inherit your plan’s limits. Default ceilings are published in the docs, and we will raise them on request for legitimate workloads — just ask before you launch, not during.
We may throttle or pause a key that endangers platform stability. When we do, you get a 429 with a retry window rather than a silent drop, and any credits held by an interrupted job are released back to you.
Suspension and termination
You can close your account whenever you like. Closing it stops future billing, keeps your vault exportable for 90 days, and then removes it permanently.
We may suspend an account for unpaid invoices, for a content violation, for suspected payment fraud, or where we are legally required to. Except for severe safety breaches, we tell you why, and you can appeal to appeals@ockeeper.com — a person reads every appeal.
Warranties and liability
The service is provided as it is. Generative models are probabilistic: they misspell text, invent hands, and occasionally refuse things they should not. We do not warrant that any specific prompt will produce any specific image.
To the extent the law allows, our total liability for any claim is limited to what you paid us in the twelve months before the claim. We are not liable for lost profits, for renders you did not export, or for third-party claims arising from images you chose to publish. Nothing here limits liability for fraud, death, or personal injury caused by our negligence, or any other liability that cannot be excluded by law.
Changes, disputes and governing law
We may update these terms. Material changes are announced in the product and by email at least 30 days before they take effect, and the version number at the top of this page always tells you which text you agreed to.
These terms are governed by Portuguese law. Before going to court, write to legal@ockeeper.com — we commit to a good-faith reply within 15 working days. If that fails, the courts of Lisbon have jurisdiction, without affecting your right as a consumer to bring a claim where you live.
Something here unclear?
Write to us and a person answers within 1 business day. We publish a plain-language changelog for every revision of this document.
legal@ockeeper.comockeeper Labs, Unipessoal Lda. · Rua da Prata 80, 1100-420 Lisbon, Portugal · replies within 1 business day
