Reflection vs prediction

Almost every editorial choice on TarotAxis follows from one position: the cards are an instrument for reflection, not a forecasting device. This is not a marketing line; it changes what we ask the model to say, what topics we refuse, how we phrase keywords, and how we frame the meaning of any individual draw. The page below sets out the reasoning so you can evaluate it on its merits.

The two traditions, briefly

Tarot has been read in both modes for most of its history. From the late 18th century onwards, French occultists like Etteilla and the Marseilles cartomants used the deck for predictive divination — timing, outcomes, names, dates. Their tradition is still alive in modern fortune-telling practice. Alongside it, and especially after the publication of the Rider-Waite-Smith deck in 1909 and the psychological writings of Carl Jung in the 1930s and 40s, a second tradition emerged: tarot as a structured prompt for introspection, with the cards functioning as archetypal symbols of the inner life. Both traditions still exist, and both have skilled practitioners. TarotAxis sits squarely in the second.

Why we chose the reflective frame

Three reasons. The first is epistemic. We have no good evidence that any deck of cards is capable of forecasting specific future events, and a great deal of evidence that humans are extremely good at retrofitting ambiguous statements to whatever happens next (the "Barnum effect"). To stake a service on predictive claims we cannot back up would be intellectually dishonest.

The second is ethical. Predictive readings are particularly easy to misuse around vulnerable people — the recently bereaved, the anxious, the chronically ill — and we did not want to build a product that could push someone towards a decision they would not otherwise make on the basis of cards drawn at random. The reflective frame still helps the same people, but it returns agency to them.

The third is practical. The cards genuinely are very good at the reflective task. The 78-card deck encodes a remarkably complete catalogue of recognisable human situations — grief, ambition, stagnation, breakthrough, betrayal, repair — and laying down three of them in response to a question reliably surfaces angles the querent had not yet articulated. That is something readers report from a wide range of belief systems, including outright sceptics.

What that means in a reading

Concretely, readings on TarotAxis avoid the future tense whenever the model can speak in the present. "You will receive news in three months" becomes "the deck is naming a period of waiting; what information are you actually waiting for?". "He will come back" becomes "the card invites you to look at what is unresolved between you, regardless of whether contact resumes". The shift is small in wording and large in effect: the seeker is positioned as the decision-maker, the cards as the prompt.

Some questions are not really about reflection — they are about outcome. "Will I get this job?", "Should I leave him?", "Will the chemo work?". We do not answer those questions. Instead, we reframe them: what would help you prepare regardless of the outcome, what is the part of the decision that is already yours, what is the question you would ask if the answer to this one were not available.

What the reflective frame costs

Some seekers come to tarot precisely because they want a forecast. TarotAxis is not the right service for them, and we will lose those users to other apps. We think that is the correct trade. The reflective frame is the only one we can stand behind without misleading anyone, and the seekers it suits — people using tarot to think more clearly — are the ones we want to serve well.

A note on the AI

Adding a large language model to the experience changes the mechanics of a reading but not its philosophy. The model is asked to speak in the same reflective register a thoughtful human reader would use, and the same hard rules apply (no hard predictions, no medical / legal / financial / relationship-leaving advice). If you want a deeper account of how the AI is constrained and where it falls short, see the AI limitations page. For the topic-by-topic editorial rules, see the ethical guidelines.