Safety

Our crisis response protocol

My Softly maintains a protocol for detecting and responding to expressions of suicidal ideation and self-harm, published here in full in accordance with California's companion chatbot law (SB 243) and New York General Business Law Article 47.

• Detection. Every conversation is automatically screened using dedicated safety technology (OpenAI's moderation models). The screening looks for explicit expressions of suicidal intent or self-harm. It is deliberately measured: ordinary sadness, grief, and hard days are part of being human, and the app does not treat them as emergencies or build a profile of your emotional state.
• Referral. When an expression of suicidal ideation or self-harm is detected, your companion's reply acknowledges what you shared and refers you to the 988 Suicide & Crisis Lifeline (call or text 988, available 24/7). We verify that the referral is actually present in the reply before it is sent; if a reply cannot be generated, a pre-written supportive message with the referral is shown instead.
• Prevention. Companions are designed never to produce content that encourages or gives instructions for suicide or self-harm. Every companion reply is screened before it is delivered; a reply that fails screening is regenerated or replaced with a safe response.
• Human review. High-risk situations are additionally surfaced to our team for human review, so serious signals are never handled by automation alone.
• Reporting. We keep an aggregate count of crisis referrals issued so we can meet our reporting obligations to the California Office of Suicide Prevention. These reports are numbers only and never include the content of your conversations.

My Softly is a companion, not a medical or mental-health service, and it is not a substitute for professional care or emergency services.

Always clearly an AI

Your companion is an AI, not a human. We never blur that line:

• Before your first conversation, the app explains that replies are generated by artificial intelligence and asks for your explicit consent.
• An "AI" label is shown with your companion's name on the chat screen at all times.
• Where required by law, a periodic in-chat notice also reminds you that you are not communicating with a human.

Content standards and moderation

• Every exchange, in both directions, is screened by automated moderation, including photos shared in chat.
• Sexual content involving minors is refused outright, and accounts involved may be suspended and reported to the appropriate authorities.
• You can report any companion message directly in the app by pressing and holding it. Reported and flagged content may be reviewed by our staff for safety, as described in our Privacy Policy.
• Accounts that misuse the service can be suspended, and specific content can be removed from the service by our team.

For adults only

My Softly is designed for adults 18 and over. The app does not accept an under-18 date of birth, and we are adopting Apple's Declared Age Range capability so that the age signal from the App Store is respected on-device as well.

Questions or concerns

If you have a safety concern about My Softly, or feedback on anything described here, we want to hear it: hello@mysoftly.ai.