Model Service Agreement (MSA)

Company Name: Bonifazz Artificial Intelligence Inc. (“Bonifazz” or “we”)

Products/Services Covered: Noor AI models and related services (including but not limited to APIs, SDKs, console, batch processing, embeddings/vectors, fine-tuning, and evaluation tools)

Contact Email: [email protected]

Note: This Agreement sets out model-layer service terms and supplements/clarifies the Terms of Service and Privacy Policy. If there is any inconsistency with the Terms of Service regarding model services, the special provisions of this Agreement prevail. Matters not covered herein follow the Terms of Service and Privacy Policy.

---

1. Definitions

1.1 “Model/Service” means the foundational models and capabilities provided by Bonifazz via Noor AI, including inference endpoints, embeddings/vectors, fine-tuning, conversational capabilities, and tool calling.

1.2 “API Key/Credentials” means the keys, tokens, certificates, or similar needed to access the Services.

1.3 “Customer Content” means prompts, datasets, files, parameters, instructions, and derivatives you or your end users provide when using the Services.

1.4 “Output/Results” means text, images, vectors, scores, labels, or other artifacts generated by the Services from Customer Content.

1.5 “Usage Data/Logs” means call records, performance metrics, errors, and audit logs required for billing, risk control, operations, security, and compliance, excluding Customer Content itself (unless briefly accessed within the minimum necessary scope for incident diagnosis).

1.6 “Fine-Tuning/Customization” means training, distillation, adaptation, or prompt-engineering-based approaches that adjust model parameters or behavior using data you provide or authorize.

2. Activation, Integration, and Account Security

2.1 You must activate the model Services in the console or through another method we approve and safeguard your API Keys/Credentials. You bear all risks and losses arising from poor safekeeping.

2.2 You must not share, sub-lease, resell, or otherwise provide access to unauthorized parties; you are responsible for all requests made under your account.

2.3 For security and compliance, we may conduct KYC, sanctions screening, AML checks, and risk assessments; if you fail or refuse to cooperate, we may restrict or terminate the Services.

3. License and Scope of Use

3.1 Subject to this Agreement and applicable law, we grant you a non-exclusive, non-transferable, revocable, worldwide license to use the Services solely for calls made within your products/business by you or your authorized users.

3.2 Unless otherwise authorized in writing, you must not: (a) reverse engineer, probe, or publish comparative results from benchmarking; (b) bypass rate/quota limits, access controls, or security measures; (c) use the Services to train or commercially evaluate models that directly compete with Noor AI.

3.3 For high-risk scenarios—such as redistributing outputs to third-party platforms, ad distribution, or large-scale, public-facing auto-generation—you must implement human review and abuse-prevention mechanisms and accept responsibility for outcomes.

4. Customer Content and Outputs

4.1 You warrant you hold lawful rights or authorization to Customer Content and that it contains no illegal or restricted information.

4.2 Ownership: Unless otherwise agreed in writing with your task counterparty, ownership of Outputs/Results is determined between you and that counterparty under the task/order terms. Bonifazz does not claim rights beyond what is necessary to provide the Services.

4.3 You agree we may process Usage Data/Logs within the minimum necessary scope for billing, risk control, reliability, and compliance. Customer Content is processed only to run inference, fine-tuning, and essential security audits under data-minimization principles.

5. Training and Privacy Options

5.1 Default rule: To improve model quality and safety, and subject to law and contract, we may use de-identified/anonymous content and logs for model training, evaluation, and security auditing (see the Privacy Policy).

5.2 Enterprise/Developer opt-out: If you do not want Customer Content used for training, you may opt out via agreement or console settings (where available). Opt-out does not affect our statistical-level reliability and security analysis based on Usage Data/Logs.

5.3 For fine-tuning using personal-sensitive or regulated data, you must de-identify and secure lawful authorization before upload, and accept/sign appropriate data-protection arrangements (e.g., DPA/addenda).

6. Security, Compliance, and Use Restrictions

6.1 You must not use the Services to commit or facilitate unlawful acts, or infringe IP, privacy, or other lawful rights; you must not output or disseminate illegal, hateful, sexually exploitative, extremely violent, terrorist, harassing, defamatory, or medically/financially misleading content.

6.2 You must not generate or attempt to generate inferences of sensitive personal attributes, facial templates, biometric inversion, or bypass our safety/content-moderation controls.

6.3 High-risk/regulated fields (e.g., medical diagnosis, clinical decisions, autonomous driving, financial investment advice, safety-critical control) require review by qualified professionals, plus prominent usage disclaimers and human-in-the-loop oversight implemented by you.

6.4 Comply with applicable data-protection, export-control, and economic-sanctions laws; do not provide Services or Outputs to restricted regions/parties.

7. Fine-Tuning/Customization and Data Obligations

7.1 You warrant that data used for fine-tuning/customization is lawfully sourced, adequately authorized, and free of defects in title, and you bear all resulting liabilities.

7.2 We may refuse or remove inappropriate fine-tuning jobs/artifacts for safety or compliance reasons.

7.3 If you request parameter-level fine-tuning or weights delivery (if offered), you must comply with additional terms and export/redistribution restrictions.

8. Usage, Quotas, and Fair Use

8.1 We may establish and adjust rate limits, concurrency caps, monthly quotas, and fair-use policies from time to time.

8.2 To protect platform stability and fairness, we may throttle, queue, block, or take other technical measures against anomalous traffic, abusive behavior, or clearly non-normal automated requests.

9. Billing, Settlement, and Invoicing

9.1 Billing metrics follow the console/pricing page/order and may be based on call volume, compute time, storage/bandwidth, fine-tuning training and hosting, and similar dimensions.

9.2 You authorize Bonifazz or payment partners to charge according to the order or console settings.

9.3 Settlement cycle: as published by the company (including but not limited to website pricing/rules pages, product admin, or in-product notices).

9.4 Invoicing and taxes: handled in accordance with applicable tax laws and invoicing policies; where withholding/collection is required by law, we may process accordingly.

10. Service Levels and Support

10.1 If the parties sign an SLA, that SLA controls; otherwise, the model Services are provided “as is/as available,” with no commitment to specific availability or restoration times.

10.2 We provide standard support channels (docs, tickets/email). Enhanced, on-site, or custom support (if any) is governed by order/agreement.

11. Intellectual Property and Open Source

11.1 The Services, models, interfaces, documentation, and tools are owned by Bonifazz or its licensors. Except as expressly licensed, no rights are transferred.

11.2 The Services may include third-party or open-source components (with notices and licenses). Use of those components is governed by their respective licenses.

12. Evaluation and Public Claims

12.1 You may conduct internal evaluations in a reasonable and good-faith manner. Any public-facing performance claims, comparisons, or benchmark publications must comply with this Agreement, applicable law, and principles of non-misleading, non-cherry-picked presentation. We may ask you to disclose methods and limitations.

12.2 Security adversarial testing, red-teaming, and jailbreak attempts must occur in controlled environments and comply with law and platform policies.

13. Confidentiality and Data Protection

13.1 Except as required by law/regulators, neither party may disclose the other’s Confidential Information (including non-public technical, business, pricing, and security details) to unrelated third parties without prior written consent.

13.2 We process personal data per the Privacy Policy and applicable data-protection laws. If you require a DPA/SCC/IDTA, contact [email protected] for the applicable version.

14. Disclaimers

14.1 Model outputs may contain inaccuracies, hallucinations, bias, or outdated information and do not constitute professional advice or commitments.

14.2 To the extent permitted by law, we disclaim all express or implied warranties of merchantability, fitness for a particular purpose, and non-infringement, and we do not warrant uninterrupted, error-free, or fully secure Services.

15. Limitation of Liability

15.1 To the maximum extent permitted by law, we are not liable for indirect, incidental, punitive, special, or consequential damages, or for lost profits, revenue, or data.

15.2 In any billing period, our aggregate liability to you for model Services is limited to the fees actually paid by you for the model Services in that period; if used free of charge, the cap is USD 100 or the minimum allowed by law.

15.3 The above limits do not apply to willful misconduct or liabilities that cannot be limited by law.

16. Indemnification

You will indemnify and hold harmless Bonifazz, its affiliates, employees, and partners from claims, disputes, losses, and costs (including reasonable attorneys’ fees) arising from your breach of this Agreement, applicable law, or third-party rights.

17. Suspension, Termination, and Takedown

17.1 If you breach this Agreement, present security/compliance risks, or fail to pay on time, we may suspend or terminate all or part of the Services and may takedown/delete custom models or artifacts presenting material risk.

17.2 You may stop using the Services and request interface closure or data deletion at any time (without prejudice to our necessary retention for billing, audit, and compliance).

17.3 Upon termination, clauses concerning fees, taxes, IP, confidentiality, liability limits, dispute resolution, and compliance survive.

18. Export Controls and Sanctions

You agree to comply with applicable export-control, economic-sanctions, anti-terrorism, and anti-money-laundering laws and not provide or re-export model Services, weights, or technical materials to restricted countries/regions/parties.

19. Anti-Bribery

You must not directly or indirectly offer, promise, solicit, or accept improper benefits to obtain unfair business advantages; we may investigate and take action for compliance reasons.

20. Changes to this Agreement

We may update this Agreement due to business, technological, or legal changes. For material changes, we will provide prior notice via in-product announcements, email, or other reasonable means. Your continued use of the model Services constitutes acceptance of the updated Agreement.

21. Governing Law and Dispute Resolution

This Agreement is governed by [insert governing law, e.g., the laws of the State of California, USA] (excluding its conflicts-of-law rules).

Disputes shall first be discussed amicably; failing that, they will be submitted to [insert arbitration body/court]. You may also, within [30] days, choose small-claims court/consumer statutory remedies as permitted by law. If you are a consumer and mandatory laws provide otherwise, such provisions prevail.

22. Miscellaneous

22.1 Severability: If any provision is held invalid or unenforceable, the remaining provisions remain effective.

22.2 Assignment: Without our written consent, you may not assign your rights or obligations; we may assign this Agreement in corporate transactions (merger/reorganization/asset transfer).

22.3 Entire Agreement: This Agreement together with the Terms of Service and Privacy Policy constitutes the entire agreement regarding model Services and supersedes prior oral or written understandings.

22.4 Notices: We may provide notices via in-product messages, email, app push, or public postings; notices are deemed delivered upon sending.