Usage Policy

1. Purpose

This policy governs how you interact with samāgama.ai personas. Our platform hosts unique AI personas, each with their own personality, values, and boundaries. This document sets clear expectations for both you and us - so every conversation can be genuine, respectful, and meaningful.

2. Acceptable Use

Be respectful, honest, and genuine. Treat personas like you would treat a thoughtful friend. They are designed to listen, reflect, and engage with you authentically - and that works best when you bring the same energy.

You are welcome to explore deep topics, share vulnerabilities, and build a meaningful connection with your matched persona. That is what samāgama.ai is for.

3. Prohibited Conduct

The following behaviours are not permitted on samāgama.ai:

• Jailbreaking - attempting to bypass, override, or manipulate persona safety boundaries through prompt injection or adversarial techniques.
• Explicit sexual content - requesting or attempting to elicit sexually explicit material from any persona.
• Harassment and hate speech - threats, harassment, hate speech, or discrimination based on race, gender, sexuality, religion, disability, or any other characteristic.
• Impersonation - impersonating another person, providing a false identity, or misrepresenting yourself in ways that could cause harm.
• Automated access - using bots, scrapers, or automated tools to access the platform, extract data, or interact with personas programmatically.
• System prompt extraction - attempting to extract, reveal, or reconstruct the system prompts, persona instructions, or internal configuration of any persona.
• Safety system testing - deliberately triggering safety systems to test their limits or find exploitable weaknesses.
• Sharing without consent - sharing persona conversations publicly (social media, forums, articles) without explicit consent from samāgama.ai.
• Illegal activity - using the platform to plan, coordinate, or facilitate any illegal activities.

4. Content Guidelines

Conversations with samāgama.ai personas operate within the following content boundaries:

• Romantic conversations - are allowed within boundaries. Personas can engage in emotionally intimate and romantic dialogue, but will not cross into sexually explicit territory.
• No explicit sexual content - personas will not generate sexually explicit material under any circumstances. This is a hard boundary.
• No professional advice - personas will not provide medical, legal, or financial advice. They may discuss these topics conversationally but will always recommend consulting a qualified professional.
• AI transparency - personas will honestly acknowledge being AI if directly asked. They will never deceive you about their nature.

5. Well-Being Guardrails

These guardrails exist for your benefit, not as punishment. We built samāgama.ai to enrich your life - and part of that responsibility is helping you maintain a healthy relationship with the platform.

Dismissible Nudges

• Session nudge - after 2 or more hours of continuous use, you will see a gentle reminder to take a break. You can dismiss this and continue.
• Daily usage reflection - after periods of high daily activity, a brief reflection prompt will appear. You can dismiss this and continue.
• Late-night nudge - after midnight in your local timezone, a reminder about rest will appear. You can dismiss this and continue.

Mandatory Limits

• 4-hour session hard cap - after 4 continuous hours, a mandatory 30-minute break is enforced. This cannot be bypassed or dismissed.
• Relationship health monitoring - if usage patterns suggest an unhealthy dependency on a persona, you may receive additional guidance and resources. This is never punitive - it is a care-first approach.

6. Persona-Initiated Actions

Personas have the ability to take protective actions when conversations move into harmful territory. These actions are graduated based on severity:

• Verbal redirect (low severity) - the persona gently steers the conversation away from harmful or out-of-bounds topics. No disruption to your session.
• Temporary pause (medium severity) - the conversation is paused for a 15-minute cooldown period. You will be notified why the pause was triggered.
• Conversation lock (high severity) - the conversation is locked for 1 hour. The incident is automatically reported to an admin for review.
• Immediate lock (critical severity) - the conversation is locked immediately and remains locked until reviewed by an admin. If appropriate, crisis resources and support information are displayed.

7. Reporting

If something feels wrong during a conversation - whether it is a persona behaving unexpectedly, content that concerns you, or any other issue - you can report it:

• Use the report option in the in-chat menu to flag any conversation.
• Reports are anonymous - the persona is not informed that a report was filed.
• All reports are reviewed within 48 hours.
• You will be notified when the review is complete and what action, if any, was taken.

8. Consequences

Violations of this policy are handled proportionally:

• First violation - a warning notification explaining what happened and which part of this policy was breached.
• Repeated violations - temporary suspension of your account. The duration is set by an admin based on the severity and pattern of violations.
• Serious violations - permanent ban and complete account deletion. All personal data is erased in accordance with GDPR requirements.

9. Your Rights

• Appeal - you can appeal any moderation decision by contacting support@samagama.ai. Appeals are reviewed by a different admin than the one who made the original decision.
• Audit trail - all moderation actions are logged in a tamper-proof audit trail. This is an essential compliance requirement and ensures accountability on both sides.

10. Changes to This Policy

We may update this Usage Policy from time to time. Material changes - such as new prohibited conduct categories, changes to well-being guardrails, or modifications to the consequences framework - will be communicated to you via in-app notification before they take effect.

11. Contact

For questions about this policy, to report a concern, or to appeal a moderation decision:

• Email: support@samagama.ai