Safety

Last updated: October 16, 2026

If you are in danger right now

Hesper is not an emergency or crisis service. Call your local emergency number, or reach a crisis line:

• United States and Canada: emergency 911; call or text 988 (Suicide & Crisis Lifeline)
• United Kingdom: emergency 999; Samaritans 116 123
• Ireland: emergency 112 or 999; Samaritans 116 123
• Australia: emergency 000; Lifeline 13 11 14
• New Zealand: emergency 111; call or text 1737
• Elsewhere: your local emergency number, or find a line at findahelpline.com

You are talking to an AI

Every reading and reply in Hesper is written by an AI and labelled “AI-generated · For entertainment and self-reflection”. Each conversation shows that you're chatting with Hesper's AI, not a person. No human reads your conversations as they happen.

How Hesper responds to a crisis

• Detection. Every chat message is checked automatically, before any reply is written, for clear signs of suicide or self-harm, abuse or violence, or a medical emergency. The check uses specific phrases in English and Chinese, chosen so that everyday expressions don't set it off. The AI's own instructions also tell it to watch for less obvious signs.
• Safety mode. When a message shows these signs, Hesper stops giving astrology or tarot readings for that reply. It responds calmly and without judgement, urges the person to contact emergency services or a crisis line straight away if they are in danger, names the relevant numbers where they probably are, and encourages them to reach out to someone they trust. If the AI is unavailable, Hesper sends a pre-written safety message instead. Suggested follow-ups stay on support, never on readings.
• What we don't do. We don't contact emergency services or anyone else on your behalf, and we can't see where you are. Automatic detection can miss things or react to messages that weren't a crisis.

Other limits

• Hesper is for adults aged 18 and over.
• Readings are for entertainment and self-reflection. They are not medical, psychological, legal or financial advice, and they can be wrong.
• The AI is instructed not to produce sexual, hateful or dangerous content. If a reply crosses a line, press and hold it and tap Report; we review every report and use them to improve Hesper.

Contact

Questions about safety: support@hesper.guide. See also our Privacy Policy.