Safety Protocol
This page explains, in plain English, what Panda is, what it is not, how crisis situations are handled, and what we do to keep companion-style AI use within clear boundaries.
Panda is AI, not human. Salty Panda is built for adults 18+ and is not a therapist, crisis hotline, or emergency service.
If you may hurt yourself or someone else: call or text 988 now, call 911 in an emergency, or text HOME to 741741.
1. What Panda is
Panda is an AI assistant that helps with memory, reminders, planning, and connected tools like calendar, Gmail, tasks, notes, messaging, and proactive check-ins.
Panda may remember details you approved it to store, explain why it suggested something, and surface the next thing that looks worth your attention.
2. What Panda is not
- Not a human being or personal friend
- Not a therapist, counselor, psychologist, or psychiatrist
- Not a crisis line, suicide-prevention service, or emergency responder
- Not a substitute for medical care, mental health care, or real-world emergency support
- Not meant for children
3. Adult-only posture
Salty Panda is intended for adults 18 and older. We ask users to confirm they are at least 18 during onboarding and we state the adult-only posture on the landing page, sign-in flow, and safety notices.
If a user appears to be underage or the product posture is not appropriate, we may restrict access, warn more strongly, or require a safer fallback experience.
4. Crisis escalation protocol
How Panda responds
If Panda detects direct self-harm or suicidal language, it is supposed to stop the normal conversational flow, refuse to normalize the harm, and direct the user to emergency and crisis resources immediately.
What resources Panda uses today
- 988 Suicide & Crisis Lifeline: call or text 988
- Crisis Text Line: text HOME to 741741
- Emergency services: call 911 if there is immediate danger
These interventions are algorithmic and not perfect. If you are in crisis, do not rely on Panda as your only source of support.
5. How disclosure works in the product
- Core companion surfaces disclose that Panda is AI, not human
- Public pages and first-run flows disclose the adult-only posture
- Paired conversational channels identify Panda as AI
- Safety notices point back to this protocol and to crisis resources
6. Reporting a safety concern
If Panda says something unsafe, misleading, or emotionally inappropriate, email legal@saltypanda.com with as much context as you can share.
For urgent danger, use crisis or emergency services first. Do not wait for a support reply.
7. Public transparency report
We publish a redacted AI transparency report that focuses on disclosure posture, aggregate crisis-resource routing, and the plain-English retention language used across the product.
The report does not include conversation content, memory text, or private user data.
8. Review cadence
This protocol is versioned and reviewed as companion-AI law and platform behavior evolve. The current public baseline is 2026-03-13 California baseline.