Load the batch.
Get a signed green OK.
The value of Serial Alice is not the dashboard — it is replacing months of manual audit work with one call. Point a period of certificates at a framework and get a single, deterministic, cryptographically-signed verdict an ESG or compliance tool can act on without a human in the loop.
An auditor's question is not "is this one certificate valid?" — it is "does this whole quarter of
evidence conform to EU AI Act / CSRD / our insurer's risk policy?" Serial Alice answers it in one
request: POST /v2/audit/conformance runs the trust engine over every certificate in the
period and folds the results into a verdict.
- A certificate conforms iff the policy finds no violation.
- The batch is CONFORMANT iff every evaluated certificate conforms — and at least one was. An empty period proves nothing and is never a clean pass.
- If it fails, you get exactly which certificates failed and why, machine-readable, ready to remediate.
The verdict
verdictCONFORMANT / NON_CONFORMANT
frameworkThe exact policy + its hash it was judged against
evaluated · passed · failedPer-certificate counts
certificates[]Per-cert pass/fail with violation reasons
coverageWhat fraction of activity was attested (anti-cherry-pick)
signatureEd25519 + public key + the SHA-256 it signed
Coverage travels with the verdict on purpose. A clean verdict over a cherry-picked slice is flagged, not laundered: the audit cannot be gamed by attesting only the favourable windows.
Verify the verdict, don't trust it
Re-check any attestation against our published key — no account, no API call.
See a real proof →
The moat: one engine, a bar per sector
What counts as proof differs by who is asking. The same cross_checked certificate
conforms to an insurer's data-centre risk policy yet is non-conformant
under EU AI Act high-risk, which demands hardware_attested. Serial Alice ships curated,
per-sector trust policies — regulatory (EU AI Act, CSRD, SOC 2, sovereign-EU) and sector (insurance
data-centre risk, GHG Scope 3 supplier evidence) — on the same engine. That parameterisation is the
barrier a generic dashboard cannot replicate.
One stack, three artefacts
The conformance verdict is one of three connected outputs — the same signed evidence, shaped for three different readers. Pick the one your counterparty needs.
Per period · this page
Conformance verdict
A signed CONFORMANT / NON_CONFORMANT verdict over a whole batch against a framework or sector policy — the auditor's automatic green OK.
Per certificate
Energy Passport →
Turn any certificate into a public, QR-scannable passport: energy, carbon, signature and on-chain anchor, verified by anyone with no account.
For your ESG tool
ESRS E1 · XBRL export →
The same evidence as EFRAG ESRS E1 XBRL — ingestible by Workiva, SAP SFM or Sweep with no manual re-keying. Automated audit, end to end.
Verifiable, not just signed
The verdict carries the public key and the exact hash it was signed over. Submit it to
POST /v2/audit/conformance/verify (no auth) and the signature is checked against the
published Serial Alice key for its kid — not the key embedded in the
document — and the verdict is re-derived from the signed counts. Alter any field and integrity breaks.
If Serial Alice disappeared tomorrow, every verdict would still verify offline in ~15 lines of Python.
Verify one yourself
Paste a conformance attestation (the JSON from POST /v2/audit/conformance), or load a
signed sample, and check it live. Nothing is stored — the check runs against
/v2/audit/conformance/verify and anchors on our published key.
What this is not
It is not legal assurance. It states, verifiably, that a period of evidence meets a named technical policy — the primary evidence an auditor samples, not the auditor's signature. A CONFORMANT verdict over low coverage is surfaced, never hidden.