Terms of Service — Nbility AI Gateway

Acceptance

By registering for, signing in to, accessing, or using Nbility services, you confirm that you have read, understood, and agreed to these terms and related policies. If you use the service on behalf of an organization, you must have authority to accept these terms for that organization and ensure the organization complies with them.

Account

You must provide accurate, complete, and current registration information, and protect your account, password, login credentials, API keys, tokens, and two-factor authentication information. Operations through your account or credentials are treated as your operations. You are responsible for risks and losses caused by leaking, sharing, transferring, or failing to secure credentials.

Service

Nbility provides an AI gateway, console, API routing, model aggregation, quota management, billing, logs, async tasks, Playground, prompt library, and related value-added capabilities. Specific service scope, model availability, rate limits, context length, pricing, regions, and feature switches follow the product pages, public notices, console display, or business contracts.

User compliance obligations

You agree to comply with laws, regulatory rules, and industry norms related to personal information protection, privacy, data compliance, cybersecurity, intellectual property, export control, content safety, and AI. You are responsible for the lawful basis and permissions needed for content you input, upload, call, generate, store, distribute, or use.

Do not submit, generate, distribute, or assist in generating illegal, infringing, fraudulent, harassing, hateful, violent, sexual, child-harming, public-safety-risk, or otherwise improper content.
Before processing another person's personal information, trade secrets, copyrighted materials, sensitive data, or enterprise data, obtain necessary authorization and meet notice, consent, security, and cross-border compliance obligations.
When required by laws, regulations, or regulators, keep necessary operation records, authorization evidence, content review records, and data source explanations.
Review model outputs with human judgment and professional care. Do not use outputs as the sole basis for high-risk decisions in medical, legal, financial, education, hiring, law enforcement, or similar scenarios.

Prohibited conduct

You may not abuse APIs, bypass quota or risk-control limits, bulk-register accounts, resell accounts or quota, attack or scan systems, interfere with services, reverse engineer, scrape non-public data, impersonate others, conceal illegal sources, use the service to evade upstream model policies, or otherwise harm the lawful rights of Nbility, users, upstream providers, or third parties.

AI content and outputs

AI outputs are generated automatically by models based on inputs and context, and may be inaccurate, incomplete, outdated, biased, repetitive, or unsuitable for a specific purpose. You are responsible for determining whether outputs are usable, and for publishing, using, modifying, training on, commercializing, or relying on them. Unless separately agreed in writing, Nbility does not guarantee uniqueness, copyrightability, legal suitability, or fitness for a particular business purpose.

AI content acceptable use policy (AUP)

This policy forms part of these terms and applies to prompts, uploaded materials, and generated text, images, audio, and video. The content standards and prohibitions above apply to both inputs and outputs. Child sexual exploitation or abuse material, terrorist content, non-consensual intimate imagery, and impersonation or deepfakes used for fraud, harassment, or infringement are prohibited. You must not bypass safety controls or assist others in doing so.

Content moderation

Nbility combines the gateway's built-in keyword filter, upstream model safety mechanisms, and human review. Automated coverage depends on the enabled rules, request type, and model provider; it is not a guarantee that every input and output is checked or that all violations are detected.

Input checks: where keyword filtering is enabled and supported, text in requests is checked against configured terms and matching requests are blocked. Upstream providers may apply additional safety checks to prompts and uploaded materials.
Output checks: we respect upstream safety refusals and filtering results. Suspected violations in generated text, images, audio, or video that are reported or otherwise identified are reviewed by staff using the available content and evidence; confirmed violations are restricted or removed within our control.
Usage checks: request logs, rate limits, and abnormal calling patterns may be used to identify repeated abuse or attempts to bypass restrictions.
Human review: support staff and personnel responsible for content safety examine reported cases and identified safety concerns, verify context and evidence, and decide on action under this policy. Access to report materials is limited to personnel who need it for handling the case and who must maintain confidentiality.
Review and improvement: we periodically review handled reports, appeals, and available risk records, and use necessary samples to assess missed violations and incorrect decisions and improve the rules. This review uses only information lawfully available for that purpose and follows the privacy policy.

Risk levels and enforcement

We assess severity, potential harm, evidence, and repeated violations to determine review priority and proportionate action. Urgent protective restrictions may be applied before human review is complete.

L1 / Critical: child sexual exploitation or abuse, terrorism, or an imminent threat to personal safety. Once identified, immediately restrict the relevant content or access and prioritize human confirmation; confirmed severe violations may result in account termination, preservation of necessary evidence, and reporting to the competent authorities as required by law.
L2 / High: serious sexual, violent, hateful, fraudulent, or infringing content. Prioritize human review and, where necessary, remove or block content and suspend related access or accounts.
L3 / Medium: other clear violations of these content standards. Conduct human review, warn the user, and remove or restrict the content; repeated violations may lead to stronger measures.
L4 / Low: minor or potentially unintentional violations. Review following a report or discovery, issue a reminder or warning, and require correction; escalate if the issue persists.

Reporting channels and principles

Anyone may report content or conduct suspected of violating this policy, including unsafe AI outputs, child safety concerns, terrorism, impersonation, deepfake abuse, or infringement of their rights.

For reports or questions, email support@nbility.ai or submit a ticket through Ticket Support (https://nbility.ai/console/tickets). If you cannot sign in, use email. For child safety concerns, terrorism, or an imminent threat to personal safety, mark the subject as urgent and briefly describe the risk.

Include the content link or request/task ID, time, relevant model or account information if known, the reason for the report, necessary evidence, and a way to contact you. Rights complaints should include proof of ownership or authorization. Do not send passwords, full API keys, unrelated personal data, or redistribute illegal material; provide identifiers or a description where possible.

Confidentiality: reporter identities and evidence are kept confidential and disclosed only as necessary to handle the case or meet legal obligations. Reporter contact details are not disclosed to the reported user without consent, unless legally required.
Impartiality: every report receives an independent, objective review based on evidence and context. A report alone does not establish a violation.
Abuse prevention: deliberately false or malicious reports are recorded and may lead to warnings or account restrictions. Good-faith reports are not penalized merely because a violation cannot be substantiated.

Report review, feedback, and appeals

After receiving a report, we record the issue, check whether the information is sufficient, and assign a risk level. Staff examine available inputs, outputs, context, and relevant records, and may request additional evidence or clarification from the parties or the model provider. We then decide whether a violation occurred, take appropriate action, and record the reason and outcome. Where contact details are available, we provide the reporter and affected user with appropriate feedback, subject to privacy, security, and legal restrictions.

L1 concerns receive immediate protective action upon identification and priority review; L2 concerns receive expedited review; L3 and L4 reports are handled according to risk and order of receipt. When acknowledging a report, we communicate the next steps and expected feedback time through the original channel. If more evidence, provider assistance, or a legal investigation is needed, we explain the delay and update the expected timeline. You may request progress through the same ticket or email thread.

If you disagree with a decision, reply to the original ticket or email support@nbility.ai, quoting the case reference and providing your reasons and any additional evidence. We will reassess the decision objectively, correct errors where found, and communicate the result. Necessary temporary restrictions may remain in place during review.

Data and privacy commitments

We process personal information and request data according to the privacy policy and applicable law. Without explicit written user permission, we will not use user inputs, personal information, or business data to train our own large models or create training datasets. Except to provide services, perform contracts, respond to lawful regulatory requirements, or act under authorization, we will not copy, retain, provide, or share such data in ways unrelated to the service.

Payment

Some features are available only to paying users. Pricing, quota, deduction multipliers, plan benefits, validity periods, invoices, refunds, and settlement methods follow the purchase pages, console display, public notices, or business contracts. We may adjust pricing or billing rules due to upstream model pricing, exchange rates, channel costs, or product strategy, and will notify users through reasonable means.

Suspension and termination

If we find account security risks, overdue payments, abnormal calls, illegal activity, infringement complaints, regulatory requirements, upstream blocking risks, or violations of these terms, we may, depending on risk level, issue reminders, restrict models, limit rates, suspend services, freeze accounts, remove illegal content, preserve evidence, terminate cooperation, or report to relevant authorities.

Liability

To the extent permitted by applicable law, Nbility will use commercially reasonable efforts to maintain service stability and security, but assumes no unlimited liability for losses caused by force majeure, network failures, upstream model or payment channel fluctuations, third-party service failures, regulatory policy changes, user misconfiguration, credential leakage, or inherent limitations of model outputs.

Breach liability

If you violate these terms, the privacy policy, applicable laws and regulations, regulatory requirements, or upstream provider policies, causing complaints, penalties, claims, losses, investigations, or additional costs to Nbility, other users, upstream providers, or third parties, you must bear corresponding responsibility and compensate reasonable losses and expenses. Nbility may immediately terminate the relevant agreement and cooperation.

Dispute resolution

Disputes arising from these terms shall be resolved by friendly negotiation first; if unresolved, may be submitted to a court or arbitration body with jurisdiction.