Methodology
How the Stack Audit receipt is generated
Eonic turns an entered supplement scenario into a deterministic Supplement Decision Receipt. The current public build uses versioned contracts, committed public rules, and explicit label-trust boundaries rather than a chatbot verdict.
First-alpha work is still staged. Public users currently receive receipts from the public seed release. Hosted reviewed releases are defined in the server contract but are not active for public traffic until a separate activation review.
Entered Scenario
The Stack Audit starts with facts supplied in the browser: products, serving details, intended intake, relationship to the product, visible label facts, selected outcomes, and optional context terms. The receipt reflects that entered scenario. It does not know what is not entered.
Manual label entries default to user transcriptions, and the builder lets users mark a fact as a machine candidate or unresolved when it is not ready for dose math. The photo-assisted preview uses a local image preview and deterministic synthetic suggestions to demonstrate review boundaries; the selected image is not uploaded in this build and the suggestions are not extracted from it.
Brand and product names entered by a user are part of the scenario text. The public repository does not publish a named-product catalog, and the current public receipt does not verify the named product against a reviewed formulation database.
Trust and Review Boundaries
Label facts carry explicit trust states. In the current public builder, user-transcribed facts can be eligible for math when the rest of the fields are computable. Machine candidates and unresolved facts are kept out of dose math until reviewed. Facts with unknown dose basis, hidden blend details, unsupported units, or missing quantities remain visible as review items instead of being silently inferred.
Review by a nutritionist or doctor can add discipline to Eonic-authored rules, references, and copy, but it does not turn a receipt into medical advice or a certification of a product. The deterministic release still has to explain the public rule, dose math, assumptions, and reference metadata used for the specific scenario.
Deterministic Evaluation
The audit API canonicalizes the submitted input, normalizes eligible label facts, computes scenario totals, applies the committed public seed rules, and returns deterministic receipt decisions, release information, and fingerprinted fields. The UI renders the engine's decisions, assumptions, included facts, excluded facts, and context limitations without doing separate dose arithmetic in the browser.
Public References
Public reference drawers show the exact public references returned with the receipt release. A public reference is not a product endorsement, not a conclusion about every possible stack, and not a substitute for the entered label facts. The current public data includes a small seed set intended to prove the receipt shape and review boundary.
Limits
Eonic does not verify product authenticity, purity, contamination, legality, manufacturing quality, or whether a label is complete. The receipt is not a diagnosis, treatment plan, or instruction to change prescribed care. It is a structured readout of the entered scenario and the public release available in this version.
Versioning
Results include a public release ID, public projection hash, exact public references used by the receipt, versioned provenance, and a deterministic fingerprint over the evaluated receipt fields. Metadata such as generation time is excluded from the deterministic fingerprint. When rules, normalization, release manifests, or contracts change, the reviewed version changes with them.