Acceptable Use Policy

1. Who this applies to

This Acceptable Use Policy applies to everyone who accesses Opus Rank, on every plan. It forms part of our Terms of Service, and the definitions there apply here.

It covers the material you supply, the content you generate, the way you interact with our systems, and what you do with the output after it leaves us. Breaching it may lead to restriction, suspension or termination, as described at the end of this page.

You are responsible for the use of your account, including by anyone you allow to use it.

2. Content you may not create or process

You must not use the Service to generate, store, transmit or publish content that:

• Is unlawful where you are, where we are, or where you publish it, or that promotes or facilitates unlawful activity.
• Sexually exploits or endangers children in any way. This is absolute. We report suspected child sexual abuse material to the National Center for Missing & Exploited Children and to law enforcement, terminate the account immediately and without notice, and preserve records.
• Infringes intellectual property — copyright, trade mark, patent, trade secret or moral rights. See our DMCA Policy.
• Violates someone’s privacy or publicity rights, including publishing personal information about a person without a lawful basis, or impersonating a real person or organisation.
• Harasses, threatens, bullies or defames any person, or incites others to do so.
• Attacks people on the basis of a protected characteristic — race, ethnicity, national origin, religion, disability, sex, gender identity, sexual orientation, age or serious disease — or promotes a violent or hateful ideology.
• Incites or glorifies violence, terrorism or self-harm, or provides instructions for producing weapons, explosives or other means of causing serious harm.
• Is sexual material, pornographic, or designed to be sexually gratifying.
• Is deliberately false and presented as fact in order to deceive — fabricated news, invented research, fake reviews, fake credentials, or fabricated quotations attributed to real people.
• Facilitates fraud — phishing pages, fake storefronts, investment or cryptocurrency scams, or content designed to obtain money or credentials by deception.
• Provides regulated advice as though it were professional advice — medical, legal, financial or similar — without the qualifications, disclaimers and review your jurisdiction requires.
• Contains malware or links to it, or is designed to compromise a reader’s device or accounts.

These rules apply to what you put into the Service as much as to what comes out of it. Prompting a model to produce prohibited material is a breach even if the model refuses.

3. Search-engine abuse

You must not use the Service to produce or publish:

• Content at scale whose primary purpose is to manipulate search rankings rather than to help the person reading it. This is the general rule, and the specific items below are examples of it rather than an exhaustive list.
• Doorway pages — large numbers of near-identical pages targeting keyword variations that funnel visitors to the same destination.
• Content farms and auto-generated site networks built to monetise search traffic with material nobody asked for.
• Scraped, spun or thinly rewritten content presented as original — including running someone else’s article through a rewrite in order to republish it.
• Parasite SEO — publishing on a third party’s high-authority domain, without their genuine editorial involvement, to borrow its ranking.
• Cloaking or deceptive redirects — showing search engines something different from what a visitor sees.
• Link schemes — content produced primarily to place manipulative links, including paid guest-post networks and private blog networks.
• Expired-domain abuse — buying a domain for its residual authority and filling it with generated content unrelated to its history.
• Publishing to a website you do not own or are not authorised to publish to.

Using the Service to draft articles for clients, for several sites you operate, or on a regular schedule, is entirely legitimate and is what the automation features are for. The distinction is not volume in itself — it is whether each page is meant to be read by a person and is reviewed by one before it goes live.

Search engines set their own rules and enforce them against your site, not against us. Complying with them is your responsibility. We mention them here because a customer whose site is penalised has usually not read them, and because our own limits are not a substitute for them.

4. Responsible use of generated output

When you publish content produced by the Service, you must not:

• Publish it without review. Language models produce confident errors. Every statistic, quotation, citation, price, date and name in a draft is your responsibility once it is live.
• Present unverified output as verified fact, particularly on health, legal, financial, safety or election topics, where being wrong causes harm to the person who believed you.
• Attribute it to a real person who did not write it, or use it to impersonate an individual or organisation.
• Ignore disclosure obligations that apply to AI-assisted content in your jurisdiction, your industry, or on the platform you publish to.
• Rely on the built-in quality checks as verification. They are automated heuristics: they miss real problems and flag correct content, and a clean report is not evidence that a draft is true.
• Assume output is original. We do not screen generated content against published material, and no plagiarism or duplicate-content detection runs in this product. If originality matters to your use, check it yourself with a tool built for that.

5. Generated images

Image prompts are checked before they are sent to a provider. Refused prompts cost you nothing — no credits are charged when a prompt is declined.

A prompt is refused outright if it seeks:

• Sexual content or nudity.
• Graphic violence — gore, mutilation, torture.
• Anything touching child safety — any prompt combining minors with sexual or suggestive content.
• Hateful imagery — the symbols and iconography of violent and hateful movements.
• Depictions of illegal activity — weapon and explosive manufacture, drug production, counterfeiting.

Two categories are adjusted rather than refused, because they create rights problems customers rarely intend:

• Recognisable brands. A prompt mentioning logos or named brands has "no real-world brand logos, trademarks or product packaging" added to it.
• Identifiable people. A prompt seeking a celebrity or a named individual has "no recognisable real people" added to it.

Every image prompt also carries a standing instruction to avoid text, letters, logos and watermarks, because models render text badly and a garbled word in a header image is worse than no word.

These checks are a floor, not a ceiling. They are pattern matching over the words in a prompt, and passing them is not permission — you remain responsible for the images you publish, including for artefacts and unintended resemblance to existing work.

6. How you may not use our systems

You must not:

• Access anything you are not authorised to access, or attempt to. Probing for other accounts’ data, tampering with identifiers in a request, or exploiting a flaw to reach content that is not yours is a breach and may be a criminal offence.
• Circumvent limits or entitlements — rate limits, concurrency limits, plan capabilities, credit costs, or the metering itself. This includes creating multiple accounts to claim the free credit grant repeatedly.
• Share or resell your account credentials, or allow anyone outside your organisation to use your account.
• Resell or white-label the Service as though it were your own product, or provide it to third parties as a service, without our written agreement. Using it to produce work for your own clients is fine; reselling access to it is not.
• Scrape, crawl or bulk-extract any part of the Service other than your own content through the features we provide for that.
• Reverse engineer, decompile or disassemble the Service, or attempt to derive its source code, models or prompts, except to the extent that restriction is unenforceable by law.
• Use the Service to build a competing product, or to train or fine-tune a machine learning model on its output.
• Interfere with the Service’s operation — denial-of-service attacks, flooding the job queue, deliberately triggering failures, or any action that degrades it for others.
• Test security without authorisation. Vulnerability research is welcome, but scan and test only your own account and tell us first. Automated scanning of our infrastructure without agreement is not authorised.
• Introduce malicious code of any kind.
• Misrepresent your identity or affiliation, or evade a suspension by creating a new account.
• Use the Service where you are legally barred from doing so, including under applicable sanctions and export-control laws.

7. Limits and fair use

The Service applies rate limits and concurrency limits per account and per plan, covering generation, rewrites, image generation, publishing, exports and authentication attempts. They exist to keep the Service available and its costs predictable.

We do not publish the exact thresholds. Publishing them tells someone intent on abuse precisely what to stay under, which makes the limits less effective at the only job they have. If you hit one in ordinary use, the application tells you and tells you when to try again.

If your legitimate usage genuinely needs more headroom than your plan allows, email [email protected] and we will work something out. We would much rather do that than throttle a customer who is using the product exactly as intended.

8. Connected publishing platforms

• Connect only sites you own or are authorised to publish to. We may require you to prove control of a domain before publishing to it.
• Use credentials scoped to what publishing needs. An application password limited to authoring is safer for you than a full administrator account, and we do not need more than the former.
• Keep your platform’s own rules. Publishing to a third-party platform is subject to that platform’s terms, and breaching them is between you and them.
• Do not use the publishing features to flood a platform or to post to sites belonging to other people.

Credentials you give us are encrypted before storage and decrypted only at the moment of a publish you initiated or scheduled. You can disconnect a site at any time.

9. How we enforce this

We investigate reports and signals of abuse. What we do depends on what happened, how serious it is, and whether it has happened before.

We do not always start at the first step. Content involving child safety, credible threats of violence, or active attacks on our systems results in immediate suspension or termination, and referral to law enforcement where appropriate.

If you believe we have acted wrongly, reply to the notice we sent you or email [email protected]. A person will look at it again. Every enforcement action is recorded with a written reason, so there is always something specific to discuss.

10. Reporting abuse

To report a breach of this policy, email [email protected] with as much detail as you can: what you saw, where, when, and any URLs. For copyright specifically, follow the process in our DMCA Policy, which sets out what a notice must contain.

We investigate every report. We may not be able to tell you the outcome, because what we do about another person’s account is generally between us and them, but the report is read.

If the content you are reporting is on a customer’s own website rather than on our systems, please read the first section of our DMCA Policy — we explain there why we cannot remove it and who can.

11. Changes to this policy

We may update this policy as the Service changes and as new forms of abuse emerge. The version and dates at the top reflect the current text, and material changes are notified in the way our Terms of Service describes.

Nothing in this policy limits any other right or remedy available to us under the Terms of Service or at law.