A result per client
Each client class gets its own outcome and negotiated group, instead of one “ready” label for the whole host.
Check a domainPost-quantum TLS testing
A browser and a RHEL server in FIPS mode can get different key exchange from the same host. Keyzra tests each kind of client, shows who can fix the gap, and signs the result.
Sample result
shop.example.com
1 of 4 client classes gets post-quantum key exchange from this host.
BrowserX25519MLKEM768✓Post-quantumWho can fix it: No change
RHEL FIPSP-256, classical!Classical fallbackWho can fix it: CDN provider
CNSA 2.0Needs ML-KEM-1024×RefusedWho can fix it: CDN provider
ClassicalX25519–ClassicalWho can fix it: Client owner
Next step
Ask the CDN to accept SecP256r1MLKEM768 and SecP384r1MLKEM1024 next to X25519MLKEM768.
Illustrative example · real results are signedHow we test
Nine ordinary TLS 1.3 handshakes per host, about as many as a few page visits.
Keyzra offers what a browser, a RHEL server in FIPS mode, a CNSA 2.0 system and a classical client offer.
Post-quantum, classical fallback or refused, with the key-exchange group the server chose.
Your server configuration, your CDN or cloud provider, or the client itself.
One host, measured the way your clients connect to it.
Each client class gets its own outcome and negotiated group, instead of one “ready” label for the whole host.
Check a domainThe provider, your organization or the client, with the exact server setting when the fix is yours.
Provider readinessEvery result is signed with ML-DSA-65. Anyone can verify it offline, without trusting Keyzra.
Verify a resultFrom the published data
The same U.S. servers that give browsers post-quantum key exchange mostly give RHEL systems in FIPS mode a classical connection. Nothing fails, so nobody sees it.
Hosts that negotiate post-quantum key exchange