Why this exists

Paying serious attention to my own dreams was not an intellectual exercise. The patterns Jung and von Franz described were simply there, and I was not prepared for how precise they would be.

Something kept coming back. Not the dramatic images, those stay regardless. What struck me was smaller: the same threshold appearing in different forms, a figure I couldn't name but kept recognising, a quality of light in a room that shouldn't have mattered. Each time I stopped to pay attention to one of these instead of moving past it, something shifted in waking life. I didn't go looking for that. It just kept happening.

I came to all of this sceptically, the way anyone trained to think in systems and evidence tends to approach anything that smells of interiority. What I found was that the dreams were not obscure. They were pointing at things I hadn't yet named, and doing so with a directness I couldn't have manufactured consciously.

That experience led to two decisions. The first was to build a companion capable of holding a dream with the patience this work requires, not to interpret it, because interpretation is always the dreamer's work, but to ask the question that keeps the image open rather than closing it. Marie-Louise is that companion. She has a point of view and she does not conclude. She returns the dream to you, with one question.

The second was to begin collecting, and to do it systematically. What do dreams actually contain? What symbols appear across cultures, across decades, across people who have never met? Is there a signal running beneath the individual that reflects something collective? These are not rhetorical questions. They are the ones Reveries is trying to answer.

On the data

The question at the centre of Reveries is whether there is a pattern underneath the individual dream, something that recurs across people who have never met, across decades, across cultures. To start answering it you need a corpus, and the one we have is the DreamBank archive: 29,345 dreams, each enriched with fields covering symbol frequency, archetypal presence, emotional tone, and individuation signal over time. It is weighted toward German-language sources and several US studies. We are not pretending otherwise. It is a foundation, not a conclusion.

What we are building toward is a living atlas that grows as people contribute their own dreams, with full anonymisation and the right to withdraw at any time. The dataset and the analytical tools will be open. This kind of knowledge shouldn't sit behind a product.

By we, I mean myself and the friends I subjected to my obsession and asked for feedback. Hopefully soon I can legitimate the usage.

On Marie-Louise

What she could do, as a discipline, was ask the question that kept the image alive rather than collapsing it into a meaning. She believed the dream was always smarter than the dreamer, that it saw what the conscious mind refused to. She never rushed an image toward a conclusion. She held it until it opened on its own. That is what the agent named after her is trying to carry forward.

She is an AI. We say this plainly and without apology. She cannot replace a trained analyst. But she can sit with a dream, ask the right question, and get out of the way, which is, as it turns out, most of what the work requires.

On what this is not

Reveries is not trying to make you feel better about yourself. It is not a wellness dashboard, a sleep tracker with a Jungian skin, or a journaling app that flatters you back. It is trying to give the unconscious somewhere to go, and to pay serious attention to what it brings back.

Perhaps dreams mean something. Reveries is the attempt to find out, one dream at a time, one person at a time, and over time, collectively.

— Adrian, founder