Live status.
Phase 1: the gateway runs as a single auditable service. Its uptime below is recorded — sampled on a fixed interval and measured over real windows, not a fabricated number. The nodes are the regions and tiers we're building toward; per-node attested hardware arrives in Phase 2.
Degraded performance
3 of 4 planned regions online · 1 degraded
- Gateway
- green
- Database
- 35ms
- Uptime
- 2d 7h
- Region
- eu-central
- TEE (planned)
- sgx
- Tier
- confidential
- State
- online
- Region
- eu-west
- TEE (planned)
- sgx
- Tier
- standard
- State
- degraded
- Region
- us-east
- TEE (planned)
- sev
- Tier
- confidential
- State
- online
- Region
- ap-southeast
- TEE (planned)
- nvidia-cc
- Tier
- sovereign
- State
- online
Incidents — last 30 days
- Confidential tier unavailable (attestation measurement mismatch)7/2/2026, 10:30:00 PM → 7/6/2026, 8:33:00 PM
Our TEE backend provider redeployed the confidential workload, changing its attested measurement. The gateway fail-closed as designed — confidential requests were refused rather than silently downgraded to a non-attested path — but the outage wasn't surfaced on this page. Resolved by verifying the new build's attestation and re-pinning the measurement. Follow-up shipped: attestation health is now part of /v1/status and this page.
Confidential tier only — standard-tier inference was unaffected.