Proof Explorer
A public explorer for inference receipts. Look up any receipt by hash and verify the model, input, and output bindings plus the TEE attestation — the receipt-of-record for audited AI.
When AI output drives a financial, legal, or medical decision, "the model said so" is not enough. Proof Explorer is the public, receipt-of-record explorer: paste a hash, verify what ran.
Look up a receipt
Every receipt has a hash. Enter it and Proof Explorer shows — and independently checks — its bindings:
| Field | What it proves |
|---|---|
| Model hash | Which model ran (not a label you trust) |
| Input hash | The input was bound to this run |
| Output hash | This output, and no other, was produced |
| TEE attestation | It ran inside a genuine enclave |
| Prev-proof hash | Its place in a chained workflow |
Auditable after the fact
Verification re-checks the proof — it does not re-run the model and does not expose the underlying data. An auditor confirms the computation happened as claimed without ever seeing the prompt or the output plaintext.
You don't need Proof Explorer to trust a receipt — its checks are the same constant-size proof
verification any device can run via omyra.proof.verify(). The explorer just makes it a URL.