LumaArc sealLUMENCOREPublic evidence surface
LumenCore public evidence surface

One proof path. One bounded decision.

LumenCore turns an AI or infrastructure claim into a reviewer-readable Proof Capsule: named source, incumbent baseline, metric locked before scoring, labeled run type, hashes, result, negative findings, claim boundary, and next pilot decision.

Evidence boundary: internal simulation, replay, benchmark, and hash records are not external validation. No membership, endorsement, field savings, deployment, certification, patent outcome, or universal model superiority is implied.

1. Inspect the canonical capsule

See what was tested, the baseline, the prelocked metric, the preserved negative result, and the exact limitation.

Open public capsule

2. Verify the artifact

Run the standard-library verifier and focused tests from a clean checkout. The verifier checks structure and hashes; it does not independently reproduce the private replay.

Run the five-minute quick start

3. Make one decision

A qualified reviewer can reject, request information, rerun under agreed conditions, authorize an external validation, or scope a paid pilot.

Current gate: external review, not field validation

Controlled external-validation gate

  1. Confirm data rights and the authoritative source.
  2. Name the incumbent comparator.
  3. Lock the metric, threshold, window, exclusions, and failure rules before scoring.
  4. Retain positive, neutral, incomplete, and adverse findings.
  5. Issue a hash-verified report with one bounded go/no-go decision.
Return to LumenCore Read the claim boundary Request a bounded review

This page is intentionally static so the public review path remains available even when an application gateway is unavailable.