The open standard for telling real from fake.
R-EAL™ — Reader Evaluation of Authenticity & Legitimacy — measures the credibility of content the way people actually establish it, and rests on one principle: credibility is earned in the open, by readers — not decreed by a platform or a machine.
R-EAL: Rigorous, Resilient, Repeatable, Revealing — and real.
R-EAL is a published open standard (v1.0), released on the International Day of Charity under the same royalty-free patent pledge as F-ACT and intended for the same neutral Swiss foundation, being established. It is free to read, implement and build on — and still open to your review as it evolves.
Open by default — free to read, free to cite, free to implement. No licence, no registration, no permission required.
The problem
Real or fake, with no common measure
Synthetic text, images, audio and video are now cheap, abundant and convincing. Provenance signals can tell a reader where a piece of content came from — but not whether to believe it. And every platform reaches its own private verdict, in its own silo, that cannot travel anywhere else. A reader is left to judge alone, with no portable, common measure of credibility to lean on. F-ACT and T-RUE keep agents accountable. R-EAL asks the human question the machines leave behind: is this real?
Generative tools make a convincing fake as cheap as a real one. Volume alone now overwhelms the reflexes people once used to tell the two apart.
Each platform judges credibility in its own silo. A verdict earned on one site cannot be carried to another, so every reader starts from zero, everywhere.
One open, portable measure of credibility — produced by readers, weighted for agreement across them — that any person or platform can generate, carry and check.
R-EAL · Reader Evaluation of Authenticity & Legitimacy
The four movements
A credibility verdict that cannot be trusted, cannot survive manipulation, cannot be reproduced or cannot be understood is a verdict in name only. R-EAL turns each word of its name into a checkable property of the verdict.
Rigorous
Is the verdict weighted for agreement across readers who usually disagree — not a popularity poll or a raw count of votes?
Resilient
Does it hold up against brigading, bot swarms and coordinated manipulation — or can a motivated crowd simply capture it?
Repeatable
Given the same evidence, does anyone re-deriving the verdict reach the same result — and can it travel with the content as a portable credential?
Revealing
Can a reader grasp the verdict at a glance — and, if they want, see exactly why it was reached?
Together they resolve on the name itself: a verdict is worth the word real only when it is rigorous enough to trust, resilient enough to survive a crowd that would game it, repeatable by anyone, and revealing enough that a reader can see why.
One Foundation, three standards
One family, three questions
F-ACT and T-RUE keep agents accountable — first the single agent, then the chains of trust between them. R-EAL turns to people: it helps a reader tell real from fake. Three standards, one Foundation, one royalty-free pledge — each answering a different question.
“Can this agent be trusted to act?”
Conformance and trust for a single agent — Grant, Usage, Audit, Revocation, Data — graded across four tiers. The question it settles: true or false?
Read F-ACT“Is that trust still good?”
Revocation across chains of agents — so a trust withdrawn upstream falls everywhere downstream, in measured time. The question it settles: still true?
Explore T-RUE“Is this content real or fake?”
Content credibility judged by readers, weighted for agreement, portable across platforms. The question it settles: real or fake?
You are hereThe two agent standards read as two sides of one coin; R-EAL is the human-facing sibling — different subject, different palette, same family and the same open guarantee.
The context
Why a reader standard, and why now
Two things have arrived at once: content that is trivially easy to fake, and a proven way for readers to judge it together. R-EAL sits where they meet.
Content Credentials (C2PA) can establish where a piece of content came from and how it was edited — but not whether people found it credible. R-EAL is designed to sit on top: C2PA tells you where content came from; R-EAL tells you whether readers found it real.
C2PA · Coalition for Content Provenance and AuthenticityBridge-weighted crowd rating — giving weight to a verdict only when people who normally disagree arrive at it together — has been shown to resist capture where raw voting fails. R-EAL takes that principle and makes it an open, portable standard rather than one platform’s private feature.
Held to the same guarantee as F-ACT.
R-EAL is developed under the same royalty-free patent pledge as F-ACT, and its core claims are intended for the same neutral Swiss foundation — so that no single company, including Society OS, can ever hold a veto over how content credibility is measured. The Foundation does not yet exist as a legal entity, and the transfer has not yet completed. Until it does, Society OS Pty Ltd makes the pledge as current owner, and that pledge binds every successor and assign.
Read the stewardship and IP-transfer storyPublished, and runnable
Run it now
R-EAL is published — and you can run it right now. There is an example credibility credential, an in-browser checker and builder, a canonical test-vector suite, a dependency-free validator, and the standard in machine-readable form. Everything checks the E1 · Declared floor for real.
Paste a credential into the checker, or build one field by field with the wizard. Grab the example credential and the validator script.
Open the kitA canonical suite of valid and deliberately invalid credentials with expected outcomes. Run them in your browser or use them as fixtures.
Run the vectorsThe standard as spec.json and spec.txt — CORS-open, built from the same source as this page so it can never drift.
View spec.jsonHelp shape what comes next
R-EAL is published, and open on purpose. If you build, moderate or simply read on the open web, the question of what is real is yours too — and the standard is better for your review. Join the working group forming around the standards family, or start with its sibling standards.
Meet OCCY
Eight arms, each cradling a node of governance. OCCY is the face of the Open Conformance Coalition — the vendor-neutral body that stewards R-EAL alongside its sibling standards. Conformance is a web, not a monolith; holding it together takes reach, not force.
Meet the Coalition