Trumis / Status
Service status
Checked live from your browser against the health endpoint — what you see is the service as reachable from where you are, right now.
Checking…
Contacting /api/v1/health.
What's monitored
| Component | Check |
|---|---|
Conversation API (/api/v1/companion) |
The health endpoint runs in the same functions environment — if health is up, turns are being served. |
| Embeddable widget & app | Static assets on Vercel's edge network — served independently of the API, so a degraded API still loads the widget, which retries gracefully. |
| AI provider | Not on this page's check. If the AI layer is unavailable the service
degrades by design: the deterministic rules engine keeps the
conversation going (method: "rules"). |
The availability model, honestly
- Trumis encrypts and retains every processed intake turn, so there is no stateful tier to fail over — availability is effectively Vercel's Sydney functions plus the edge network.
- A person mid-conversation loses nothing in an outage: the transcript lives on their device, and the next successful turn continues where they were.
- Historical uptime reporting from an independent monitor will be published here as it accrues; licensed agreements can include specific availability commitments.
- Incidents worth knowing about are noted on this page and in the changelog.
Monitor it yourself: GET https://trumis.com.au/api/v1/health —
documented at developers.