How a Session Unfolds
You can speak or type your question. Speech is first converted to text while limited cues such as speaking rate, pauses, energy, and emotional distribution are extracted. The system may ask up to three rounds of clarification before choosing an appropriate structure from the three-card “Inner Mirror,” five-card “Relationship Mirror,” or five-card “Two-Path Choice” spread.
The system randomly selects cards and their upright or reversed orientation. During interpretation, it combines the question, card positions, cards already drawn, and your feedback into a narrative. The final report brings symbolic language back to small, actionable steps rather than making definite predictions.
What Role Do Voice Cues Play?
Voice only provides supporting context about expressive state. For example, more pauses or lower energy may suggest “perhaps you are hesitating.” It cannot read minds or prove hidden motives, personality, or mental illness from speech. The system should use uncertain wording such as “may,” “seems,” and “might be worth noticing.”
The Boundaries of Randomness and AI
What Is Retained
Temporary audio is deleted after voice analysis. The session is used to continue the current flow, expires after about 24 hours, and can also be deleted manually. Do not include ID numbers, home addresses, account details, or other sensitive information in your question.