Trust & limitations
What this proves — and what it does not.
Truth Engine is an evaluation system. Confidence should come from observable behavior and explicit boundaries, not inflated claims.
Current environment
What you are testing.
A real engine-backed evaluation path with a harmless synthetic provider. Public access is protected; the underlying Runtime remains private.
The public interface reaches the real evaluation Runtime through the protected gateway.
No production payment, order, banking, health, or customer system is involved.
Public Playground access is gated and rate-controlled before it reaches the private service.
The current public environment is for evaluation and testing only.
Observed behavior
The flagship case.
One public evaluation demonstrates a specific difficult case. It does not establish a universal provider guarantee.
The provider may have acted while the application lacks enough evidence to claim occurrence or non-occurrence.
A retry request is evaluated, but no second consequential effect is sent without authority.
The approved reconciliation path performs read-only observation.
Truth changes only after admissible evidence supports occurrence.
What we do not claim
No inflated guarantees.
Truth Engine does not claim exactly-once behavior against arbitrary external providers.
The current public environment is not a production certification.
Current public evidence does not establish production-scale performance.
The flagship case proves that one blind resend was denied in that evaluated path. It is not a guarantee across every provider or integration.
Uncertainty
Reconciliation does not mean resolution.
UNKNOWN is a legitimate truth state. A later observation can justify a change, or uncertainty can remain.
Available evidence justifies that the intended effect occurred.
Available evidence justifies that the intended effect did not occur.
Available evidence does not justify either certainty.
Disclosure boundary
Public where useful. Private where necessary.
Problem, principles, truth states, high-level diagrams, Playground, public API documentation, limitations, trust and privacy.
Redacted architecture/security material and institutional evaluation material.
Full Kernel and Runtime internals, implementation details, complete adversarial corpus, security-sensitive internals and private frozen artifacts.