mid-verify 0.2.1 checks a sign-in on the server. The crate says it does not call back.
mid-verify 0.2.1's docs say a server checks a MATA sign-in locally and does not call MATA. An October 18 note says DLDeploy uses that crate.
The short version
The crate docs say verification of a sign-in token runs locally on the relying party, with no HTTP call to MATA during the check. The October 18 Remade field note says DLDeploy signs people in through mid-verify 0.2.1 and does not fetch a JSON Web Key Set or call back to an identity provider.
The check the crate page describes
The docs.rs page for mid-verify 0.2.1 describes a relying-party verifier. Given a compact token and a small configuration, it runs cryptographic checks on the machine that received the token and returns the identifier, a roster version, and the claims the token disclosed. The page says those checks make zero HTTP requests to MATA infrastructure. Everything the cryptography needs is inside the token. Anything else is configuration the relying party already holds, or a later comparison against a version that party stored for itself.the October 18 field note.
Four checks, named at the level of the docs
The documentation lists four steps: a genesis signature checked against the key encoded in the identifier, a roster chain whose versions move only forward, a check that the signing device is in the current roster, and an outer signature check. This page names the steps because the docs name them. It does not reproduce a code sample, a clock skew, or a command. The dependency is where the check runs, which the page says is the relying party.the distributed cloud desk.
The repository's list of missing calls
The MATA mID repository says the verifier is a pure function with no input-output, dual-licensed MIT or Apache-2.0, and that a deployment does not need a client identifier, a redirect allow-list, a token endpoint, a key-set fetch, or a resolver call. The identifier, it says, is the public key. A list of calls the project says it does not make is the project's claim. A packet capture would be someone else's. Asked directly, What does the field note add that the crate page does not? That DLDeploy is the service using the crate for sign-in, and that the deployment does not fetch a key set or call an identity provider. The note is the deployment sentence. The crate page is the mechanism.Futureweb.
The field note is the deployment
Remade's learn index, dated October 18, 2026, says DLDeploy signs people in with MATA through mid-verify 0.2.1, that the check stays on the server, and that there is no key-set fetch and no call back to an identity provider. That is one house note, used here as the sentence that names a real service. The crate page is the mechanism. The index line is who says they turned it on. This brief does not paste the note's essay, because the index line and the crate docs are the documents in hand.the key stays with the holder.
A minted device name is the neighboring record
This paper already filed a device that draws its own P-256 key and derives a did:mata from that key. mid-verify is the other side of a sign-in: the server checks a token that carries that kind of identifier. The device page is about the key on the device. This page is about the check that does not phone an identity provider. They meet at the identifier and not at the product.a device that mints its own name.
What this page will not claim to have read
The October 18 headline is on the learn index. The essay URL under that headline was not the page this brief opened. Sentences about DLDeploy come from the index line. Sentences about the four checks come from the crate documentation and the repository. A later reading of the essay can add detail. It cannot be invented here. Asked directly, Which version do the docs and the field note name? 0.2.1. The crate description says it is the relying-party verifier for a mID token. The October 18 field note says DLDeploy uses that version for MATA sign-in.
What a capture would change
A log that shows the server asking MATA whether a token is valid would contradict the crate's zero-call sentence for that deployment. A version newer than 0.2.1 with a different rule would replace the version in the field note. Until then, the dependency is a local check, documented on the crate, and a service that says it uses that crate. Asked directly, What does the documentation say about network calls? Zero runtime HTTP calls to MATA infrastructure during verification. The inputs the check needs travel inside the token. Configuration and any rollback comparison stay on the relying party.
Rollback is the party's own memory
The docs say a relying party may compare the token's roster version with the newest version it has stored for that identifier, after the local check. That comparison uses the party's database. It is not a call to a directory. Storing a version is state. It is still not an identity provider. A control plane that lives in one region is a different dependency. Asked directly, What does the repository say the token replaces? A client identifier, a redirect allow-list, a token back-channel, a key-set endpoint, and a resolver call. The repository says the token carries its own resolution data and that the identifier is the public key. Those absences are the project's description.when one region is the control plane.
What stays on the record
mid-verify 0.2.1, a local verification, no runtime call to MATA during the check, a repository list of the back-channels it says it omits, and an October 18 index line that DLDeploy signs people in that way. The key the person holds is checked where the service runs. The short answer this page will stand on is this: The crate docs say verification of a sign-in token runs locally on the relying party, with no HTTP call to MATA during the check. The October 18 Remade field note says DLDeploy signs people in through mid-verify 0.2.1 and does not fetch a JSON Web Key Set or call back to an identity provider.
Sources
The reports this brief is filing. Futureweb did not republish them.
Questions
Which version do the docs and the field note name?
0.2.1. The crate description says it is the relying-party verifier for a mID token. The October 18 field note says DLDeploy uses that version for MATA sign-in.
What does the documentation say about network calls?
Zero runtime HTTP calls to MATA infrastructure during verification. The inputs the check needs travel inside the token. Configuration and any rollback comparison stay on the relying party.
What does the repository say the token replaces?
A client identifier, a redirect allow-list, a token back-channel, a key-set endpoint, and a resolver call. The repository says the token carries its own resolution data and that the identifier is the public key. Those absences are the project's description.
What does the field note add that the crate page does not?
That DLDeploy is the service using the crate for sign-in, and that the deployment does not fetch a key set or call an identity provider. The note is the deployment sentence. The crate page is the mechanism.
Does a local check mean the server never sees the token?
No. The relying party receives the token and checks it. Local means the check does not place a call to MATA to decide whether the token is good. The token still arrives at the server the person is signing in to.