DESTILL.ai
Verify a credential β
What's real and independently verifiable today β and what's still on the roadmap. We publish what we can prove and label the rest.
| Capability | Evidence |
|---|---|
| ML-DSA-65 signing + verification (FIPS 204, post-quantum) | β 25/25 LEDGER tests; in-browser verify live-confirmed |
| RFC 6962 append-only transparency log (inclusion + consistency proofs) | β exhaustively tested across tree sizes |
| RFC 3161 independent trusted timestamp | β live against public TSAs (freeTSA, DigiCert) |
| did:web issuer + browser verifier | β resolves & verifies client-side |
| AEGIS injection screen on CPU, before the model | β running; no published rate until a reproducible public rerun |
| Capability | Honest status |
|---|---|
| Hardware key custody (HSM / AWS Nitro Enclave) | π§ software custody today; keys in process memory |
| Full TSA certificate-chain verification | π§ we verify grant + digest-match + time, not yet the cert chain |
| Production did:web hosting at the live domain | π§ staging exists; not yet served at /.well-known/did.json |
| Reproducible public AEGIS benchmark | π§ rerun pending; role-play jailbreaks are the largest known gap and need a trained model |
STENO, QFVC, and QFAI-C ("compression") were removed after we measured them: STENO's benchmark was salted from its own templates; QFVC wasn't a real codec. We don't sell numbers we can't reproduce.