Why not use the standards that already exist
Anybody who knows this field asks the same question within a minute of arriving: verifiable credentials are a finished W3C standard, the European Union is deploying wallets, and Ethereum already has an identity registry. This page answers it, including the parts where those answers are better than this one.
The short version
Every standard below assumes an issuer who signs. That assumption is the whole problem. Across the 54 registers in this index, the number of issuing authorities that sign anything at all is zero.
Revow is not a competing credential format. It is what you do in the years before the issuer signs, and it is designed so that the day an issuer finally does, the earlier documents get stronger rather than being thrown away.
W3C Verifiable Credentials
The Verifiable Credentials family reached W3C Recommendation in 2025, seven specifications covering the data model, data integrity, securing credentials with JOSE and COSE, and a status list for suspension and revocation. It is mature, properly specified, and this lab has no quarrel with it.
It describes how an issuer, a holder and a verifier exchange a signed credential. It is a format and a role model. What it cannot do is produce an issuer where there is none, and the data model itself is careful about what a signature buys: “verifiability of a credential does not imply the truth of claims encoded therein”.
Where Revow differs: the subject signs, not the issuer, and the document is a claim rather than a credential. The difference is deliberate and it is stated on every page. A claim proves that a key said something. It does not prove the licence, and the protocol never pretends otherwise.
Where the standard is better: its revocation machinery is more developed than RVW-6, and its selective disclosure work solves a privacy problem Revow has not had to face yet because it carries almost no personal data.
If an authority were ready to issue signed credentials tomorrow, that standard is the one to use, and this lab would say so.
The European wallet and eIDAS 2
Regulation (EU) 2024/1183 entered into force in May 2024 and requires every member state to make a digital identity wallet available by the end of 2026, with acceptance obligations phasing in through 2027. The wallet carries electronic attestations of attributes, and professional qualifications are named among them.
The part that matters most here is the public body attestation: a document issued from an official register that carries the same legal value as the paper original. That is an authority signing a statement about a licence, mandated by law, across twenty-seven countries. It is closer to what this lab asks for than anything else in existence.
Three differences remain. It is holder-centric: the credential lives in the citizen wallet and is presented on request, so there is no file a third party can fetch, hash and keep. It is European: twenty-seven of the thirty-seven jurisdictions in this index at best, and none of the Australian, American, Asian or African ones. And it is not here yet.
There is also a genuine tension. Wallets are built for selective disclosure, so a citizen reveals the minimum. Revow asks an authority to publish a list. Those are opposite instincts, and anybody who prefers the wallet answer on privacy grounds is making a real argument, not a lazy one.
ERC-8004 and onchain identity registries
ERC-8004, Trustless Agents, was created in August 2025 and is still a draft. Its registries launched in January 2026 and now span around thirty networks, drafted by people at MetaMask, the Ethereum Foundation, Google and Coinbase. It gives an agent an identity as an ERC-721 token, plus reputation and validation registries, and reuses EIP-712 and ERC-1271 rather than inventing new cryptography.
It also includes optional domain verification through a well-known file, which is the same instinct as RVW-2 arrived at independently for a different population.
It is solving an adjacent problem: software agents with no legal existence, where there is no issuing authority to appeal to and reputation has to be built from scratch. A licensed professional is the opposite case. The authority exists, the licence is real and legally meaningful, and the problem is that the authority publishes nothing a machine can check.
One number from that ecosystem is worth carrying over. By April 2026 roughly fifteen thousand of about one hundred and sixty-five thousand registered agents were dead or uncallable. Registration is not proof, on any chain. That is the same lesson this lab learned from registers that list people who stopped practising years ago.
Decentralised identifiers
DIDs standardise the identifier layer: an identifier that resolves to key material, with a method describing how that resolution happens. They are the right abstraction if you expect many kinds of key infrastructure to coexist.
Revow deliberately does not use them. The subject is a plain address and the document is named by its own digest, so there is nothing to resolve and no method to agree on. A person checking a claim by hand needs a shell and nothing installed.
That is a trade, not a victory. Refusing the resolver layer costs portability across key systems, and a future version of this lab may be wrong about it. The open problems page records that where it belongs rather than hiding it here.
Side by side
| Question | The standards above | Revow |
|---|---|---|
| Who signs | The issuing authority | The subject, and anybody who checks |
| What it needs to start | An issuer willing to sign | Nobody |
| What a signature proves | That the issuer said it | That a key said it, dated and attributable |
| Where the document lives | With the holder, presented on request | Wherever the subject publishes it, named by its digest |
| Scope | European law, or the agent economy | Any jurisdiction, any regulated trade |
| What happens when an authority joins | It becomes possible | Existing documents reach a higher level unchanged |
The last row is the design decision everything else follows from. An authority that starts publishing later changes nothing about documents already signed. It only raises the level they reach.
Where this lab would be wrong
If the European wallet ships on time, is widely accepted, and member states issue professional qualification attestations in volume, then for twenty-seven countries the interesting problem is solved by law and Revow is at best a bridge.
That would be a good outcome and this lab would write it up in the changelog rather than arguing with it.
What would still be missing is everywhere else, and the part no wallet addresses: a client who needs to check, in the moment, that the stranger asking for a payment is the professional they claim to be, without a prior relationship and without an app both sides agreed to install.
Objections welcome, through the contact form. The ones that change a specification get published with their date.