Warranted trust — reliance you can check.
OpenReason builds open infrastructure for warranted trust — reliance grounded in evidence you can check, not in authority or reputation. Its first standard is the OpenReason Protocol. Everything we build is open, neutral, and verifiable.
Consequential decisions — in policy, in AI, in markets — increasingly turn on reasoning that is opaque: the data behind it, the assumptions inside it, the people it affects. When the reasoning can’t be seen, it can’t be challenged on substance, and disagreement collapses into who you trust. As those same decisions move onto agent-to-agent rails, the problem compounds: a machine has to decide, in milliseconds, whether to rely on a claim it cannot inspect.
OpenReason makes the reasoning an object: structured, forkable, and — since v0.3 — cryptographically verifiable, so both a person and a machine can judge how far to trust it.
Rational, Empathy-Informed Ethics
Five commitments about method, not conclusions.
Attest, never grade
ORP attests to provenance and reasoning — who said what, when, citing which sources and assumptions. It never issues a verdict on whether a claim is good or correct. The moment a standard grades content, it loses the neutrality that makes an attestation worth anything.
Check without trusting
Every claim should be checkable without trusting the author — by hand for a person, by signature for a machine.
OpenReason was founded by the Public Reason Project, an open-standards project for epistemic accountability. Its argument — that every consequential decision rests on data, and almost no one examines how that data was made — is why this protocol exists. OpenReason is the project’s first standard; the manifesto and foundations behind it live there.
OpenReason is an open protocol, not a product — anyone can implement it; no single entity owns it. The code and specification are Apache-2.0 (an explicit patent grant for a standard others implement). The governance model is designed so the project can function without its founders: disagreement is resolved by forking and showing better reasoning, not by vote.