Prove you dreamed it first.

A premonition only counts if it was written down first.

Everyone has felt it: a dream that seems to land days before the event it described. But by the time the event arrives, memory has softened, and no one can tell whether the dream truly came first or whether the mind quietly rewrote it after. That gap is where every claim of foresight collapses.

Oneiratory is built to close that gap and to answer one narrow, testable question: did the words of this dream provably exist, unchanged, before the event it seemed to foretell? Nothing more mystical than that, and nothing less rigorous. We take no position on why a dream might arrive early. We only make sure that when one does, it can be proven, and when it does not, that is on the record too.

1. Write the dream, and declare the target

   You record the dream and, in advance, state plainly what would count as it coming true. The condition is sealed alongside the dream. You are calling the shot before you take it, the same discipline a scientist uses registering a hypothesis before the experiment.

2. Seal it

   The exact text and its condition are run through SHA‑256, a standard, one-way method that turns any text into a short, unique fingerprint (the same technique that secures passwords and banking). Change a single character and the fingerprint changes completely. That fingerprint is stamped to a public timeline, so the words are now locked to this moment and cannot be moved.

3. Wait

   The dream stays private and unread while the world turns. Nothing about it can be edited, and its sealing time is fixed for good.

4. Reveal, and let a stranger verify

   When the event arrives, you unseal the dream. Anyone can recompute the fingerprint themselves and confirm the words existed, untouched, before the event. Not on our word. On the arithmetic.

Two sealed dreams. One resolved as a hit, one as a miss. Both are proven, and both are permanent. The instrument does not decide whether a dream is meaningful, only whether the call was made in advance and whether the declared condition was met.

• You declare the target before the shot. The oldest trick in prophecy is the vague dream stretched afterwards to fit whatever happened. Oneiratory forbids it: the success condition is committed up front and sealed.
• The misses are public. A ledger that only shows hits is advertising. The Registry shows the failures next to the successes, because the failures are what make the successes mean anything.
• A stranger can check it. Every proof stands on a recomputable hash and a fixed timestamp. You never have to trust Oneiratory, or the dreamer. You can verify it yourself, offline.

Scattered anecdotes have never been able to settle whether dreams can reach into what is coming. The evidence was always unprovable: a story retold after the fact, with no way to date it and no way to count the times it failed.

A sealed registry changes what is possible to study. Every entry is a pre-registered prediction with a declared condition, a fixed timestamp, and a recorded outcome, hit or miss. Kept at scale, that stops being a pile of stories and becomes a dataset: falsifiable, time-anchored, and open to anyone who wants to measure it.

For researchers of dreams, memory, and cognition, that opens real questions. How often do pre-registered dream predictions actually resolve as hits, against what chance alone would give? Do certain kinds of dream, or certain dreamers, resolve more often than others? How much of memory’s certainty about a dream is built only after the event? Oneiratory cannot answer these on its own, but it can supply the one thing this question has always lacked: honest, dated, verifiable data. Anonymised submissions may be used in research and cited in scholarly work.