A compiler for AI thought

The chain,
not the narrative,
carries the warranty.

We are handing the keys of science and engineering to autonomous AI agents, but there is a fatal flaw in the architecture: a brilliant analysis and a confident hallucination look exactly the same. Eigenius is a typed knowledge graph that records how something is known — not just what is known.

The trust gap

Every scientific claim has a reasoning chain.
Most of it is invisible.

Consider the sentence “EIG_0291 was identified as a strong CDK2 inhibitor (IC₅₀ = 85 ± 5 nM, n = 6).” Whether drafted by a human author or by an AI agent, that sentence carries five distinct implicit assertions about measurement, statistics, vocabulary translation, decision rule, and sample attribution. None are visible in the published sentence. Each is load-bearing. Each is where a hallucination or unjustified assumption can hide while looking exactly like the correct article.

The accountability surface is binary: either you trust the author, or you don’t use the result. At human authoring speed that trade-off was tractable. At LLM authoring speed it is not.

The structural answer

Every claim lands in one of four warrant categories.

The chain records how a claim is justified. Three warrant kinds are mechanically warranted — an instrument run, a deterministic computation, or a formal proof — and can be independently re-executed by anyone with read access. The fourth carries an authoritative warrant: a named declarer is on record, and the citation chain is examinable.

Mechanical warrants — re-executable
Observed Instrumental

Raw measured data with unbroken instrument provenance.

Audit fingerprint Instrument run, time, operator.

Example: Six raw IC₅₀ readings from a Kinase-Glo plate run.

Derived Computational

Outputs of deterministic computation, rerunnable byte-identically.

Audit fingerprint Procedure, inputs, kernel that ran it.

Example: A t-statistic and p-value the statistics institution emitted.

Verified Formal-proof

Claim accepted by a machine-checked mathematical proof.

Audit fingerprint Proof artifact, checker, independent re-execution.

Example: A Lean-verified regulatory primary-endpoint result.

Authoritative warrant — examinable
Declared Authoritative

Asserted by a named human, organisation, or policy document.

Audit fingerprint Declarer identity, citation, organisational authority.

Example: A 100 nM threshold cited from a methodology document.

The four categories are peers, not a hierarchy. None is more “true” than the others; they differ in what kind of warrant they carry. The chain doesn’t prevent an AI from declaring something false, but it does prevent the AI from passing a declaration off as a mechanical warrant.

Where to start

Three ways into the platform.

Watch

Three minutes on what Eigenius is.

The infrastructure-for-truth pitch in video form. Additional explainers covering specific concepts (cross-vocabulary bridges, AI reasoning verification, kernel design) are embedded on the relevant Concepts pages, and the full playlist on YouTube ↗ walks through everything.

The platform pitch: a typed knowledge graph that records how scientific claims are reached, not just what is claimed. The chain, not the narrative, carries the warranty.

Research

The case in long form.