Aletheia — AI peer review

What this is — and what it is not

Aletheia is an AI adjudication process, not a substitute for conventional peer review. It is fast, systematic and consistent: it applies the same checks to every submission and shows its work. It does not confer academic credit, it is not accredited by any body, and it does not replace review by domain experts at a journal.

We think it is useful anyway — for preprints, industry research, technical reports and work that would otherwise get no structured review at all. But we would rather tell you its limits up front than have you discover them later.

What certification actually asserts

That the submission was run through our published review procedure, that the review found its claims to be supported by the evidence provided, its methods internally consistent, and its citations real and correctly represented.

It does not assert that the work is correct. Review catches what review can catch. A certificate is a record of a process, not a guarantee of truth.

How review works

Step one

Intake

You send the manuscript and any supporting data. We confirm scope and tell you if it is something we can meaningfully review.

Step two

Adjudication

Independent automated reviewers assess claims against evidence, check internal consistency, and verify every citation resolves and says what it is cited for.

Step three

Report

You receive the full adjudication — findings, severity, and what would need to change. This happens whether the work passes or not.

Step four

Certification

Work that passes may carry the Aletheia mark, published together with its review so readers can judge the review as well as the work.

Submitting

Submissions go through the Aletheia intake portal — a short structured form, not an email. The answers decide which checks run and what the certificate can honestly claim.

Open the submission portal

What it asks

What kind of paper it is — empirical, computational, review, methods, dataset descriptor, engineering feasibility, techno-economic, position paper or case study. Different work needs different review.
What it contains — statistics, derivations, a dataset, code, human or animal subjects, physical performance claims, a cost model. Each declaration routes specific checks: say it reports statistics and every reported figure gets recomputed.
What you intend it for — internal discussion, a technical report, a preprint, a conference. This sets the scope of the certificate, so it never claims more than the work was reviewed for.

It does not ask who you are.

No author, no institution, no funder, no prior-submission history. Those are exactly the fields that bias a review, and Aletheia has no legitimate use for them — so it never collects them.

Do I keep the rights to my work?: Yes. Submitting grants us only what is needed to review it, and — if you accept certification — to publish the review alongside your work. You retain copyright and can withdraw before certification.
What if it does not pass?: You get the full report and nothing is published. We do not publish failed reviews, and we do not disclose that a submission was made.
What does it cost?: Nothing during this initial period. If we introduce a fee it will be for the review itself and charged whether or not the work passes — we will never price a favourable outcome.
Who reviews it?: Automated reviewers, with a human editor accountable for the outcome. We publish the procedure so you can see what was applied. See our methodology.
Is there a conflict of interest?: Orrery publishes its own research as well as certifying others'. We will not review work that competes with, or is authored by anyone connected to, our own publications, and any relationship is disclosed on the certificate.