Continuous probe, last 24 hours: 95 of 95 checks passed (one check per minute, measured by a separate monitoring process that keeps counting through this API's own restarts).
This page is served by the API process itself, so reaching it means the service is up. The process counter resets on every deploy; restarts are deploys, not outages, which is why the 24-hour probe figure above is the one that matters. Every figure on this page is measured, not asserted.
A 60-check audit runs against the live service every day: validator anchors, the HTTP contract (free 400 before any payment), manifest agreement across catalog/OpenAPI/x402.json, the POPI meter invariant (no identity data in logs), settle-safety on every paid route, and parity between the free browser checker and the paid API. Failures alert the operator immediately.
Latest run: 63/63 passed, 29 July 2026, 21:40 SAST.
History accumulates from 29 Jul 2026 onward; each daily audit appends a row here.
An Uptime Kuma instance probes this API every minute around the clock and alerts the operator the moment a check fails, independently of the API process itself. Full disclosure: it currently runs on the same host, so it catches process, service and tunnel failures (the common cases) but would not witness a whole-host outage. For that reason no uptime percentage is published yet; a probe in a separate failure domain comes first, because verza does not publish numbers it cannot back.
None recorded since launch (2026-07-16). When one happens it will be listed here with cause and resolution, not deleted.