AI Transparency — How Tarosyn Uses AI

This page explains the AI models Tarosyn relies on, what data we send, and the limits of AI-generated readings.

Models We Use

• GPT-4.1 Mini (OpenAI) — User-facing tarot readings, horoscopes, dream interpretations, and Oracle chat.
• GPT-4.1 Nano (OpenAI) — Background tasks, notifications, and lower-priority AI jobs.
• GPT-Image-1 (OpenAI) — AI-generated tarot card art and hero card creation.
• GPT-4o Mini Transcribe (OpenAI) — Voice input transcription (where enabled).

What Data We Send

For reading generation, Tarosyn sends: the selected tarot cards, your question (if provided), your Mystic Persona preference, and the reading spread context. We do not send your full reading history, personal messages, or birth data to OpenAI.

Limits of AI Readings

AI tarot readings are generated content for entertainment and personal reflection. They are not predictions, not professional advice, and should not be used as a basis for major life decisions. Always consult qualified professionals for medical, legal, or financial matters.

Content Safeguards

OpenAI's usage policies and Tarosyn's prompt configuration together prevent the AI from generating content that is hateful, sexually explicit, self-harm-promoting, or harmful. See our AI Safety Policy.

Responsible AI Principles · AI Bias Report · AI Training Data Policy