What we do not say
The four things NeuralRing never claims, and why each refusal is what makes the evidence usable.
Never a compliance verdict
Compliance is a determination your auditor or your regulator makes over your whole organisation. We produce the evidence about your AI supply; the judgement stays theirs.
Not where inference physically ran
We state the execution country at the assurance the evidence supports — a provider's declaration, a documented region, a probe — and we say which one it is.
Not that a request was sovereign
We create a verifiable attestation for every request: what was chosen, under which policy, at which assurance. Whether that is sovereign enough is your policy's call.
Not a claim raised by signing it
A signature proves who wrote the record and that it was not altered. It does not make a stated fact a verified one.
What we do not say.
- We do not say your AI is compliant.
- We do not say where inference physically ran.
- We do not raise a claim by signing it.
- We do not say a request was sovereign. We create a verifiable attestation of what it depended on, at the assurance the evidence supports; your policy makes the call.
If a surface ever seems to say one of these four things, that is a defect — tell us.