Does Anyone Need Agent Identity Yet?

Why this page is not the other one

The audience page asks WHO would need this and answers with hypotheses, each paired with what would kill it.

This one asks WHETHER anybody does yet, and answers with a measurement instead, because we happen to have built an instrument that would register demand and it can simply be read.

The instrument, and what it says

A credential verification through our live endpoint writes a meter event. An authorisation check writes one. Nothing else can.

Read today, that table has no rows at all. Not few. None.

So no credential this product issued has ever been verified through us, and no authorisation check has ever run.

The supply side, and an honest label on it

Twenty-one agents. One hundred and sixty-two credentials. Forty-two operators.

Those counts include our own testing and we cannot separate ours from anybody else's from these numbers, so do not read them as adoption. They establish only that the product works and has been exercised.

Put beside the meter, they say something sharper than either number alone: this product has issued one hundred and sixty-two credentials and consumed none.

Why the zero survives its own caveat

The meter is blind to the mode we recommend. Offline verification touches nothing of ours, so a relying party doing exactly what the documentation says generates no event. A zero here is a lower bound on consumption, not proof of none.

Which is why the second signal matters: there are no webhook subscriptions either. Not one.

And a webhook subscription does not touch the live path, so it is not blind in the same way. A party integrating in earnest wants to be told when a credential of theirs is revoked, and nobody has asked to be told anything.

Two independent signals, one of them without the blind spot, agreeing. That is a stronger reading than either alone, and it is why this page is willing to answer the question at all.

What the answer actually is

No, not yet, on the evidence we can see.

A product that cannot be resolved by third-party tooling and has never been audited would show this reading whether or not the underlying need existed.

So this measures us, not the market. Anyone quoting it as evidence about agent identity generally would be over-reading it, and we would rather say that than enjoy the citation.

What is real, because absences alone would misdescribe us

The product works. Credentials issue, anchor, present and verify, and the counts above are the evidence that the path is exercised rather than theoretical.

The instrument is honest. It records verification rather than registration, deduplicates against inflation, and is constrained in the schema to two event types, which is why its zero can be trusted as far as it reaches.

And the estate publishes it. A zero on your own demand meter is the least flattering number a company holds, and printing it is the whole point of this exercise.

What would move it, in the order it would move

1. One metered verification from a party unaffiliated with us. That single row changes this page's answer, and it is a much lower bar than a customer.
2. One webhook subscription, which would say somebody intends to operate rather than experiment.
3. A resolver driver, without which third-party tooling cannot resolve our identifiers at all.
4. A published external audit. None exists.