Safety and Crisis-Response Protocol

1What Mentalyze Is — and Is Not

Mentalyze is a non-clinical wellness tool that provides a space to talk through and reflect on your own thoughts through artificial intelligence. You are always interacting with an automated AI system, not a human being, and not a licensed professional.

2When the Protocol Activates

Our systems are designed to recognise language that may indicate a user is at risk, and to respond accordingly. The protocol is designed to activate when a user appears to express, among other things:

• Suicidal thoughts, intent, or planning;
• Self-harm, self-injury, or severe self-neglect;
• Disordered-eating behaviours that may pose a risk to health;
• An intent to harm another person; or
• Acute distress, or symptoms suggesting a mental-health emergency (for example, references to psychosis or detachment from reality).

3How the Service Responds

When the protocol activates, the service is designed to:

• Avoid generating any content that encourages, instructs, or facilitates self-harm or harm to others, and avoid content that normalises or reinforces such behaviour;
• Respond supportively and direct the user, clearly and promptly, to appropriate human help and crisis resources (see Section 4);
• Where appropriate, temporarily pause the ordinary conversational experience and surface crisis resources instead;
• Remind the user that they are speaking with an AI and that the service is not an emergency or clinical resource; and
• Where our systems or personnel determine that a user's needs exceed the intended scope of the service, restrict, suspend, or redirect access in the interest of the user's safety.

These measures are automated and are not a guarantee of detection. No automated system can reliably identify every situation, and users should never rely on the service in an emergency.

4Crisis Resources We Surface

When appropriate, the service directs users in the United States to resources such as:

• 988 Suicide & Crisis Lifeline — call or text 988, or chat at 988lifeline.org (24/7);
• Crisis Text Line — text HOME to 741741 (24/7);
• Emergency services — 911, for any situation involving immediate danger to life.

Users outside the United States are directed to locally appropriate emergency numbers and crisis services where available.

5AI-Status Disclosure

Consistent with applicable law, the service discloses clearly and conspicuously, within the interface, that the user is interacting with artificial intelligence and not a human. This disclosure is provided on first use, again when a user returns after a break, on request, and by way of a periodic in-session reminder.

6Access Limitations

The service is restricted to adults (18+) and is not available to residents of jurisdictions that prohibit AI-based mental-health, psychological, therapy, or counseling tools, which currently include Nevada, Illinois, Maine, and Rhode Island. Access from those jurisdictions is technically restricted.

7Transparency and Reporting

Where required by applicable law, we maintain and, on the schedule set by that law, publish aggregated, non-identifying statistics on how often the safety protocol activated and how often users were referred to crisis resources. These statistics contain no personal or conversation content. Where a state requires reports to be filed with its Attorney General (for example, beginning 1 July 2027 under Rhode Island S 2195 / H 7350), we file them as required.

8Governance and Review

This protocol is owned by the Company and is reviewed at least annually, and after any material change to the service or any relevant safety incident or legal development. Changes are reflected in an updated version of this public summary.

9Contact

Questions about this protocol may be sent to sueta.llc@gmail.com.