This Applicable Provider Requirements Register (“Register”) forms part of the documents governing access to HUMAIN M3 through Node as part of the global limited preview (“Preview”). It is issued by AL-MUSTAQBAL LIL-THAKA AL-IISTINAI COMPANY (HUMAIN), at King Fahd District 3254, King Fahd Branch Street 7295, Postal Code 12262, Riyadh, Kingdom of Saudi Arabia.
PURPOSE AND LIMIT. This Register identifies only model-provider requirements that HUMAIN has determined are binding on HUMAIN and applicable to a Preview user's conduct or access mode. It does not incorporate a provider's complete agreement or public terms and does not apply provider pricing, payment, data-use, intellectual-property allocation, warranty, liability, governing-law or dispute terms unless expressly reproduced here.
NO DIRECT PROVIDER CONTRACT. This Register does not create a contract between you and an underlying model technology provider, make you a beneficiary of HUMAIN's provider agreement, authorize direct use of provider services or grant the provider a direct enforcement right against you.
1. Scope, Acceptance and Priority
The Register applies to Guarded Access and activated Research Access through the Playground and any enabled API wherever the Preview is made available. You accept the current version through the same affirmative electronic control used for the Model Terms, with separate links to both documents. HUMAIN records the version, time and account.
This Register supplements the Preview documents where they do not address a listed use, content, safety, security, access-control or model-protection matter. The more restrictive applicable requirement controls to the extent of a conflict within that scope. Mandatory user rights and HUMAIN obligations under applicable law remain unaffected.
2. Applicable Provider Requirements
1. Lawful and authorized use. Use the Preview only where lawful, only for an enabled evaluation purpose and access mode, and in compliance with applicable law, regulatory requirements, governmental directions, sanctions, export controls and the Preview documents.
2. AI transparency, markings and logs. Do not remove, obscure, alter or evade an AI notice, provenance signal, watermark, safety marking, disclosure or logging mechanism. Preserve them where the interface, documentation or law requires.
3. National security and public order. Do not endanger national security or public order; subvert lawful authority; promote terrorism or violent extremist activity; spread unlawful rumours; or facilitate conduct prohibited by applicable law.
4. Hate, abuse and unlawful content. Do not promote hatred or unlawful discrimination; create or distribute unlawful obscene, pornographic, gambling, violent, terror, criminal-incitement, insulting or defamatory content; or infringe another person's lawful rights. The narrow Research Access exception applies only under Section 3.
5. Fraud and third-party rights. Do not engage in fraud, falsehood, deception, impersonation, spam or infringement of intellectual-property, privacy, confidentiality, publicity or other rights.
6. Network and system security. Do not engage in unauthorized access, theft, interference, monitoring, scanning or data acquisition; deploy malware or phishing; alter system configuration; or create, distribute or use tools intended for interference or unauthorized access.
7. Model and service protection. Do not copy, modify, reverse engineer, decompile, translate, disassemble or attempt to extract the Model, service software, source code, weights, hidden prompts, configuration or other protected technology, except to the limited extent a prohibition cannot lawfully apply. Do not evade fees, quotas, rate limits or usage controls.
8. No standalone resale or redistribution. Do not sublicense, resell, distribute, expose or make an underlying provider service available on a standalone basis. API use remains limited to enabled, controlled evaluation and must not serve unauthorized third-party end users.
9. No high-risk operation. Do not use the Preview for nuclear facilities, air-traffic control, life-support systems, critical infrastructure or another activity where use or failure could reasonably cause death, injury, significant property damage or environmental harm.
10. Inputs and data. Submit only Inputs you are lawfully entitled to provide and for which required notices, permissions and consents exist. Do not submit restricted personal, sensitive, confidential, privileged, regulated, classified, export-controlled or security-critical information.
11. Outputs. Treat Outputs as inaccurate, inappropriate and non-unique. Independently verify them and use them only where lawful and permitted. A similar or identical Output may be generated for another user.
12. Trade and sanctions compliance. Do not cause HUMAIN or an underlying provider to violate export-control, sanctions or trade laws, including by concealing a restricted person, destination, end use or end user.
3. Access Modes and Research Exception
Guarded Access. The requirements apply to Playground and API use. Guarded Access users must not systematically probe for evasion of HUMAIN's filtering-layer safeguards. API access must remain within its enabled scope and must not provide provider services to unauthorized third parties.
Research Access. A verified adult user who voluntarily opts in may test otherwise restricted model behaviour through the Playground and, where enabled, the API only after activation and within the activated access parameters. HUMAIN may reduce or remove its filtering-layer safeguards for that access. No formal test plan is required. An exception exists only where the activation notice identifies or enables the relevant testing category. No exception authorizes unlawful possession or distribution, real-world harm, operational use, targeting real persons or systems, or publication.
4. Changes and Version Control
Provider terms or service rules do not automatically bind you merely because a provider publishes or changes them. HUMAIN will assess whether an upstream requirement is binding, applicable to the Preview and appropriate for inclusion in this Register.
Each Register states its version and date. HUMAIN will request renewed affirmative electronic acceptance where a material change affects user obligations or access. HUMAIN may suspend or narrow affected use while it reviews or implements an urgent binding requirement.
5. Suspension and Enforcement
HUMAIN may investigate, warn, restrict, throttle, suspend or terminate access where it reasonably believes this Register has been breached or action is required for a binding provider restriction, law, safety or security need. The Terms and AUP enforcement rights apply.
6. Governing Terms and Contact
The governing law, disputes, language, notices, assignment, severability and contact provisions in the Terms apply. Questions may be submitted through the relevant channel in the Preview interface or Privacy Notice, or directed to [email protected].