Tarosyn Legal

AI Transparency Statement

Last updated May 31, 2026

Tarosyn uses artificial intelligence to power much of what makes the app meaningful — personalised readings, conversational guidance, dream interpretation, and more. We believe you should always know when AI is involved and understand exactly what it can and cannot do. This statement explains every AI touchpoint in plain language.

where ai appears in tarosyn

Tarot Readings

Every spread — from a single card to a ten-card Celtic Cross — is interpreted by an AI language model. The model weaves together card symbolism, your chosen spread type, the optional context you provide, and the Mystic Persona you've selected to produce a personalised reading.

Oracle Chat

The 1:1 Oracle is a conversational AI interface. Your messages are sent to a language model along with a persona-specific system prompt and recent conversation history. The Oracle responds in the voice of whichever Mystic Persona is active.

Mystic Personas

Eight distinct Mystic Personas shape the tone, vocabulary, and spiritual framing of every AI response. Each persona is a custom system-prompt template; the underlying model is the same language model across all of them.

Dream Interpreter

When you describe a dream, the text is sent to a language model that draws on archetypal symbolism and the context you provide to surface themes and possible meanings.

Photo Reading

If you share an image, it is sent to a multimodal language model that analyses the visual content and produces a tarot-inspired reflection based on what it perceives. Images are not stored beyond the API call.

AI-Generated Artwork

Celestial Children, Living Cards, AI Sanctuary cards, and the World Card auction use an image-generation model to create original artwork based on prompts derived from your birth data, card combinations, or deck settings.

Blog & Marketing Content

Some blog posts and marketing copy published on the platform are drafted with AI assistance. Tarosyn staff review and may edit AI-generated editorial content before publication.

which models we use

Tarosyn's AI features are powered by OpenAI APIs. The specific models we currently use are:

gpt-4.1-mini — Primary model for user-facing readings, Oracle chat, and dream/photo interpretation gpt-4.1-nano — Background tasks such as blog scheduling, marketing generation, and low-priority content jobs gpt-image-1 — AI-generated artwork for Celestial Children, Living Cards, and the World Card auction gpt-audio — Spoken-voice generation for audio readings and narrated guidance gpt-4o-mini-transcribe — Audio transcription for voice-to-text features

When no OpenAI API key is configured (for example, in a self-hosted deployment), Tarosyn falls back to built-in template responses so the app remains functional.

what data is sent to ai models

To generate a reading or Oracle response, Tarosyn sends some of your data to OpenAI. Depending on the feature, this may include:

The cards drawn and their positions in the spread Any optional context you type before starting a reading Your Oracle conversation history (recent messages only, not your full history) Your birth date and name, when you've opted in to personalised horoscopes or birth-chart readings Dream text you enter in the Dream Interpreter Images you upload for a Photo Reading Your selected Mystic Persona and Truth State dial setting

Your account credentials, payment information, and private messages to other users are never sent to AI models. OpenAI's data-handling practices are described in their Privacy Policy.

what ai can do Produce personalised, symbolically rich interpretations of tarot spreads tailored to the cards drawn and the context you share Engage in thoughtful, persona-driven conversation that supports self-reflection and exploration Surface archetypal themes and patterns in dreams, photographs, and astrological data Generate original artwork that reflects your unique combination of cards, birth data, or spiritual attributes Adapt its tone, depth, and framing to your chosen Mystic Persona and Truth State preference Provide astrological context, numerological insights, and daily guidance informed by established traditions what ai cannot do

Tarosyn's AI is a tool for personal reflection, not a professional authority. It has meaningful limitations you should be aware of:

It cannot provide medical, psychological, legal, or financial advice — do not make significant life decisions based solely on AI readings It does not predict the future with any certainty; readings are interpretive, not prophetic It can produce plausible-sounding but factually incorrect information (a known behaviour called 'hallucination') — verify important facts independently It does not remember your full conversation history between separate sessions unless you explicitly share context It cannot assess genuine mental health crises — if you or someone you know is in distress, contact a licensed professional or emergency services It has no knowledge of events that occurred after its training data cut-off date Its responses are probabilistic outputs, not deliberate human opinions — they may occasionally feel off-topic or generic safeguards & content policy

Every AI request passes through Tarosyn's server before reaching OpenAI. We apply the following safeguards:

System-level content filters instruct the model to stay within spiritual and reflective themes and to decline requests for harmful, illegal, or explicit content The Truth State dial lets you control how direct or gentle AI language is — it does not override safety filters Rate limits and token caps are enforced server-side to prevent abuse of AI generation endpoints AI-generated artwork goes through OpenAI's own safety classifiers before the image URL is returned Tarosyn staff review AI-generated blog content before publication your choices

You are never required to share personal context with AI. Every reading and Oracle session accepts an optional "context" or "intention" field — you choose how much to share.

If you prefer readings without AI interpretation, Tarosyn displays the traditional card meanings alongside every AI-generated reading so you can compare or rely on either.

For questions about how your data is handled more broadly, see our Privacy Policy. To opt out of AI features entirely or request deletion of your data, contact us at [email protected].

questions about our ai use

If you have questions or concerns about how Tarosyn uses AI, please reach out at [email protected]. We are committed to being transparent and will respond to every enquiry.

AI Safety Policy AI Bias Report Acceptable Use Policy changes to this statement

As Tarosyn's AI features evolve we will update this statement to reflect new touchpoints, model changes, or data-handling practices. Material changes will be announced via in-app notification or email.

Questions? [email protected]