Why Nobody Prices Agent Identity Yet: The Natural Unit Is the One You Cannot Meter

The event that carries the value in agent identity is verification, and verification is designed not to be observable. That is the whole argument, and everything below is why it follows rather than being a stage the market will grow out of.

What is already established, so this page does not repeat it

We publish no price and our free tier advertises more than it allows. Our code is instrumented to bill for verification, attributed to the operator. And the instrument reads zero.

Those are the measurements. This page is the one argument that follows from them.

The natural unit is verification, and verification is unobservable by design

Issuance is the wrong unit. Issuing a credential nobody checks produces no value, so charging for it prices the wrong event.

Verification is the right unit, because that is the moment a credential does work: somebody decided something on the strength of it.

And verification is offline. Checking a proof needs no call to the issuer, which is exactly the property that makes the credential worth holding. A relying party who has never heard of us can verify without telling us.

So the event that carries the value is the event the vendor cannot see. That is not an early stage of a market. It is a structural consequence of building the thing correctly, and any priced-per-check model has to either give up the offline property or accept that it meters a fraction.

And the payer is not the beneficiary

The relying party gets the assurance. The operator holds the credential.

Our own meter attributes a verification to the operator, which is a defensible choice and still leaves a two-sided market: the side that benefits is not the side invoiced.

Those markets price late, because one side has to be dense enough to make the other pay, and neither side is dense anywhere we can measure.

What is real, because absences alone would misdescribe us

The instrument exists and is honest. It counts verification rather than registration, deduplicates against inflation, and is constrained in the schema to two event types.

And the offline property is real and worth more than the metering problem it creates, which is why this page frames the tension as a cost of a good decision rather than a defect to fix quickly.

What would resolve it

1. A metered verification from an unaffiliated party, which is the first observation of the unit actually occurring.
2. A decision about the offline property, kept or traded, made deliberately rather than by default.
3. Density on one side of the market, which nothing currently indicates.
4. A published external audit, which is upstream of anybody paying for anything here.