Privacy-Adjacent Trust Layer

Evidence without exposure.

ZK-AI is a parallel accountability layer. It issues and anchors receipts about private events without requiring those events to be publicly readable.

Sealed commitments

Sensitive claim fields can be replaced by cryptographic commitments, allowing a receipt to prove structure without publishing the private values.

Evidence roots

Evidence sets can be committed through deterministic roots and disclosed later to an authorized verifier or dispute process.

Selective disclosure

A receipt holder can reveal only the commitment openings or evidence items needed for an audit, dispute, or authorized review.

Local-first tools

Proof Lab keeps image bytes in the browser. The demo signing endpoint receives only JSON metadata and the fingerprint needed to mint a demo receipt.

Public recognition is not public disclosure.

A receipt can be verified locally, and some receipts can later be recognized on ZK-AI Chain. That does not mean the underlying claim, file, payment, or private event must be published. The public layer can carry anchor facts while private evidence remains private.

Disputes need controlled evidence.

During an audit or dispute, the holder of a receipt can disclose the specific evidence needed to prove the claim. The verifier checks the disclosed material against the receipt commitments instead of trusting a platform dashboard or mutable export.

Proof Lab keeps files local.

Proof Lab is designed so image bytes do not leave the browser. When demo signing is available, the mint endpoint accepts JSON only: the fingerprint, media type, size, and a client nonce. It does not need multipart uploads, remote URL fetches, or image storage.