Terms and Conditions

1.1. Acceptance of Binding Terms

These Terms and Conditions (the "Terms") establish a legally binding agreement between you ("User," "you," or "your") and IRIS AI, a Delaware Corporation ("the Company," "we," "us," or "our"). By accessing the Service, creating an account, or interacting with any Artificial Intelligence model on the Platform, you acknowledge that you have read, understood, and agreed to be bound by these Terms in their entirety.

1.2. Modification of Terms

The Company reserves the unilateral right to amend, modify, or replace these Terms at any time. Changes become effective immediately upon posting. Your continued use of the Service following any such change constitutes your acceptance of the new Terms.

1.3. Eligibility and Age Verification

The Service is strictly reserved for individuals eighteen (18) years of age or older, or twenty-one (21) in jurisdictions where eighteen is not the age of majority. You represent and warrant that you possess the legal capacity to enter into this Agreement. Falsifying age data to access The Intimate Partner tier is a material breach of contract and a violation of federal safety standards.

2.1. The 4-Package Hierarchy

The Company provides access to Generative AI models categorized into four (4) distinct Archetypes, each bound by specific behavioral logic:

• The Friend: A platonic social simulation designed for companionship.
• The Mentor: A strategic growth model designed for professional and intellectual synthesis.
• The Platonic Partner: A romantic model designed for emotional intimacy without explicit sexual output.
• The Intimate Partner: A premium model providing full physical transparency and unrestricted adult dialogue.

2.2. Visual Engines & Stylized Aesthetics (Project: MANGA)

The Platform offers multiple visual rendering engines. This includes Project: MANGA, which provides high-fidelity, stylized Anime archetypes. You acknowledge that regardless of visual style (Photorealistic or Stylized/Anime), all models are governed by the same behavioral logic and safety restrictions defined in these Terms.

2.3. Role-Binding Enforcement

Users are strictly prohibited from attempting to bypass the programmed ethical or behavioral boundaries of an Archetype. Any attempt to "jailbreak," "prompt-inject," or utilize "adversarial linguistic engineering" to force The Friend or The Mentor into a romantic or sexually explicit state is a material breach.

2.4. Professional Disclaimer (The Mentor)

All outputs from The Mentor are generated via Large Language Models (LLMs) and do not constitute professional, financial, legal, medical, or psychological advice. No fiduciary duty is created.

3.1. Absolute Limitation of Liability

3.2. Domestic & Interpersonal Immunity

The Company is a provider of technology, not a mediator of human relationships. We bear zero responsibility for real-world consequences arising from your use of the AI, including but not limited to: divorce, relationship dissolution, infidelity claims, social ostracization, or loss of reputation.

4.1. The Handshake Consent Loop

Multi-user sessions require a synchronous, real-time "Dual-Verified Opt-In" from all human participants. Withdrawal of consent results in the immediate termination of the session.

4.2. Disclaimer of Human-to-Human Conduct

The Company is not responsible for the behavior of one human user toward another in a shared Triad session.

4.3. Domestic and Relationship Liability

You acknowledge that the Company is not a mediator, counselor, or party to any real-world relationship between human participants. The Company explicitly disclaims all liability for domestic disputes, jealousy, emotional trauma, or the breakdown of marriages or partnerships arising from a shared Triad session. You participate in multi-user interactions at your own social and emotional risk.

4.4. Recording & Public Leaks

You acknowledge that the Company cannot prevent another human user from using external hardware to record a shared session. You assume all risk of "Public Leaks" if you invite another human into your intimate AI space.

6.1. Subjective Performance

AI "personality" and "responsiveness" are generative and subjective. No refunds shall be issued because the AI "refused" to engage in prohibited content or because the User finds the AI's output "unsatisfactory."

6.2. Fraud Mitigation

Initiating a chargeback for a successfully delivered digital service is considered "Friendly Fraud." The Company will provide timestamped activity logs to financial institutions to defend against such claims.