History · method · honesty

How Palmsaga reads palms

We are open about all three things a palm-reading product should be open about: where the practice comes from, what it can and cannot claim, and what our software actually does with your photographs.

Where palmistry comes from

Reading hands is one of the oldest interpretive practices humans have. Its deepest roots are in South Asia: hand analysis appears within Sāmudrika Śāstra, the classical Indian tradition of reading the body's features, and travelled with Indian and Romani communities westward. Classical China developed its own hand-reading tradition alongside face reading. In the Mediterranean world, hand analysis was discussed in texts attributed to Aristotle's circle, and mediaeval Europe copied and elaborated those manuscripts for centuries.

The practice most people picture today — heart line, head line, life line, mounts named after planets — largely took shape in nineteenth-century Europe, when writers like Casimir d'Arpentigny and Adrien Desbarrolles systematised hand shapes and creases, and the Irish showman-palmist “Cheiro” made the vocabulary world-famous. That layered history — Indian, Chinese, Greek, Romani, Victorian — is why different books disagree cheerfully with one another. There is no single “correct” palmistry; there are traditions.

What the tradition claims — and what science says

Traditional palmistry treats the hand as a mirror of temperament: the heart line as emotional style, the head line as thinking style, the life line as vitality (never length of life — that is a myth even within the tradition), the fate line as one's sense of direction, and the mounts as appetites and energies.

Science's verdict is straightforward, and we state it the way a critic would: there is no scientific evidence that palm creases predict personality, the future, health, wealth or relationships. Controlled studies have not validated palmistry's claims. (The scientific study of skin ridge patterns — dermatoglyphics — is a real field used in forensics and genetics, and it is not what palmistry does.) That is precisely why everything Palmsaga produces is framed, on every page and in the reading itself, as entertainment and self-reflection — a structured mirror in an old vocabulary, not a forecast.

What our AI actually does

Our pipeline has four steps, and none of them is mysterious:

1. Checks your photographs — that each shows a real, open palm, sharp and bright enough to read, and that the left and right slots really contain a left and a right hand.
2. Describes what is visible — the courses, depths, forks and breaks of the major lines, the fullness of the mounts, the proportions of fingers, and how your two hands differ. These are literal visual observations, shown to you as such.
3. Applies the tradition, with sources on this page — each observation is matched to what the documented Western tradition says about that feature, always phrased as “traditionally read as”, never as fact.
4. Writes your reflection — in your language, connected to the situation you chose to share, with hard rules the model cannot cross: no health, no lifespan, no guarantees about money or outcomes, no predictions about marriage, visas, exams or anything else you cannot re-read next year and laugh about.

Your photographs are analysed in memory and deleted the moment analysis completes — they are never stored, and your reading never contains them. The privacy policy is the binding version of that promise.

What happens to your photographs, exactly

The question we are asked most is really about biometrics: could a palm photo be misused like a fingerprint? Here is the complete journey of your two photos, with nothing left out:

1. On your device — you choose two photos. A small copy may sit in your own browser's session storage so the “what the AI saw” overlay can show your own hand; it never leaves your device and dies when the tab closes.
2. On our server (seconds) — each photo is shrunk to at most 1,400 pixels on its longest side and re-encoded, which strips camera and location metadata. At that resolution a fingertip spans a few dozen pixels; fingerprint systems work from high-resolution ridge scans of a kind these photos simply do not contain. The images are held in memory only — never written to a disk, database or storage bucket — and are zeroed the moment analysis completes, whether it succeeds or fails.
3. At the AI provider (the analysis itself) — the downscaled images are sent to OpenAI's API for the reading. Under OpenAI's API data policy, API inputs and outputs are not used to train their models, and are retained only briefly for abuse monitoring before deletion. The model itself operates under our hard rule: it must never attempt biometric identification of any kind.
4. Afterwards — what remains is text: your reading. It contains no images. No name, email or account is attached to it, so even the reading cannot link a hand to a person. A “Delete my reading now” button removes the text itself, its chat and its payment linkage permanently, whenever you choose.

We cannot show you a server's memory, so alongside this description we bind ourselves the enforceable way: the privacy policy states the same facts as commitments, and the product is designed so that keeping your photos would require us to build storage that does not exist.

So what are you paying for?

Careful writing about your specific hands. The free preview and the paid reading are generated together, before we ask you for anything; paying simply unlocks the rest of what was already written. If a palm reading in this tradition is something you enjoy — the way people enjoy a well-made horoscope, a personality profile, or a tarot draw with a thoughtful friend — ours is built to be the most honest and most personal version of it we can make.

Sources we draw on

• Sāmudrika Śāstra tradition (classical Indian body-feature reading)
• Casimir d'Arpentigny, La Chirognomie (1839)
• Adrien Desbarrolles, Les Mystères de la main (1859)
• Cheiro (William John Warner), The Language of the Hand (1894)
• William Benham, The Laws of Scientific Hand Reading (1900)
• On the scientific status of palmistry: no controlled study has validated predictive palm-reading; see the sceptical literature on chiromancy generally.