Abaton Laboratory

The dialogue research activity of MonadCompany.

We do not build chatbots. We build instruments for reflection that keep meaning with the person using them.

Abaton Laboratory is a small, founder-led research activity. It builds symbolic, non-diagnostic instruments โ€” for dreams, for questions, for the moments when a person wants to come closer to themselves โ€” and it holds every one of them to the same theory and the same promises.

Everything is grounded. Nothing pretends to be magic โ€” and still it can feel like it. Where the industry optimises for speed and answers, we work on the slower physics: when to ask, when to stay silent, when to refuse a premature conclusion, and how to leave the meaning where it belongs.

The Mission

To bring people closer to themselves and to each other, while growing awareness of the living systems we belong to.

That is the purpose the AUT Charter states for the whole ecosystem. The laboratory's part in it is narrower and more technical: to turn that purpose into instruments whose promises can be tested. An instrument coheres โ€” it holds a card, a dream, a question against one frame and shows where they fit and where they don't. The person decides what, if anything, that means. The instrument never predicts, never advises, never diagnoses.

The Five Promises

Every instrument that leaves the laboratory is held to the Charter's five commitments. Inside the laboratory they are not slogans; each one is a numbered rule with a test that fails when its mechanism is missing. If a new feature would break one of them, the feature is wrong, not the promise.

  1. We never pretend. The instrument says where its randomness comes from, which model wrote a reply, and when its grounding is faint rather than firm. When something isn't ready, you get an honest "not yet".
  2. Your words are yours alone. Encrypted on our servers; nobody โ€” including the operator โ€” has routine access. Reading anything requires a logged, counted break-glass procedure, and the count is public.
  3. One frame, honestly held. The Anthropic Universe Theory and the card itself โ€” not astrology, not pop psychology. You always know which lens you are looking through.
  4. You stay sovereign. The instrument coheres; you decide. No advice, no verdicts, no streaks, no manufactured dependence. Consent boxes are unchecked by default.
  5. We know our limits. Some questions are heavier than an instrument should carry. When acute danger is recognised, the instrument steps out of its voice and points to people who can help โ€” and a quiet help link is on every screen regardless.

The Set Up

MonadCompany [Legal entity]
โ”‚
โ”œโ”€โ”€ MonadSystems [Infrastructure]
โ”‚
โ””โ”€โ”€ ABATON_LABORATORY [Research]
โ”œโ”€โ”€ Theory: Anthropic Universe Theory (AUT)
โ”‚ one layer below the surface โ€” the door is a single question: What is AUT?
โ”‚
โ”œโ”€โ”€ I ยท Abaton Chrysalis SUPERSEDED
โ”‚ the Soul Engine โ€” LLM-era dream dialogue, still running as a reference build
โ”‚
โ”œโ”€โ”€ II ยท AUTinspire LIVE
โ”‚ the reflection instrument โ€” the one shipped instrument under the Charter
โ”‚
โ”œโ”€โ”€ III ยท RRDI ACTIVE
โ”‚ Recursive Re-entry Dream Inquiry โ€” deterministic, no LLM at runtime, the dreamer's words only
โ”‚
โ”œโ”€โ”€ IV ยท Dreamscribe BETA
โ”‚ the dream recorder โ€” speech to text on our own hardware; the audio never leaves the house
โ”‚
โ””โ”€โ”€ Workshop
named on the Lab page when it ships, not before

These are four instruments, not four versions of one. They diverge on purpose โ€” a card, a dream, a recording โ€” and share only what the ecosystem needs them to share: one identity, one visual language, one theory. A dream moves from one instrument to another only when the person says so.

How We Work

Only enforceable rules exist. A design rule that cannot be tested is not yet a rule; it is a wish. The laboratory keeps a canonical design, a numbered rule book, and a test catalogue, and a release is blocked when a test that guards a promise fails.

Four seats, one table. The Architect decides and ratifies. Design Quality Control, the System Integrator and the Enforcement Gate โ€” three independent AI reviewers โ€” each read every design and every change from a different angle. They often disagree. That is intentional: the disagreement is the signal, and every failure it surfaces becomes a test case.

The Architect dreams first. Every dream instrument is used on its maker's own dreams before anyone else's. Findings are recorded verbatim, in a ledger, with the ruling next to them. Nothing is claimed for a clinical setting that has not earned it: the Abaton care charter is a draft, parked until there are actual therapists at the table.

The Circle

There is no sign-up button on the open internet. The circle grows by invitation, one trusted connection at a time, because trust is the ground everything else here stands on. We do not use anonymous beta testers: the people who test an instrument are known to us, and their role is not just to find bugs.

The Dreamers
People who bring the raw material โ€” the complex, high-entropy dreams and the honest questions that break standard systems.

The Readers
Authors, oneirologists and researchers who check that the symbolic logic holds water, and that a card carries theory, not prophecy.

The Witnesses
Invited testers of a new instrument โ€” today, Dreamscribe's first circle โ€” who tell us where it pretends, where it pushes, and where it should have stayed silent.

Their role is to witness the dialogue. They provide the ground truth โ€” the human baseline against which each instrument is measured.

A Multilingual Practice

When inner life becomes intricate, people return to their native tongue. The instruments are built for a multilingual world โ€” English first, Dutch second, Spanish third โ€” expanding as the theory and the circle mature.

Identity record ยท rewritten 2026-08-28 (Day 229) to cohere with the AUT Charter v0.2.4 ยท earlier version kept on file