Status

Live system status.

AskOrac runs on a small stack of trusted upstream providers. This page polls each of them directly and tells you the truth — whether everything's fine, or whether something in the chain is having a bad day.

All systems operational

Live status. Refreshed every 60 seconds.

Upstream providers

The services AskOrac depends on to function. Each row reads from the provider's own public status feed — click the arrow to open their full status page.

Vercel

Operational

Application hosting and edge network

All Systems Operational

Supabase

Operational

Postgres database, authentication, storage, and vector search

All Systems Operational

OpenAI

Operational

Default LLM completions and embeddings

All Systems Operational

Anthropic

Operational

Alternate LLM for BYOK customers

All Systems Operational

Also in the stack (not polled)

These providers use status page platforms with a different feed schema — we haven't added a normalizer for them yet. Click through for their live status.

Incidents on AskOrac

Problems that happened on infrastructure we operate — distinct from upstream provider issues above.

No incidents to report.

When something breaks on our end, we'll list it here with a plain-language post-mortem — what went wrong, how long it affected customers, and what we changed to keep it from happening again.

How we handle incidents

The protocol, written down.

  • Detect. Internal monitoring watches the same upstream providers listed above, plus our own API response times and error rates. We hear about problems before customers do, most of the time.
  • Communicate. When a customer-affecting issue is confirmed, we email affected accounts within 30 minutes of detection with what we know so far — not polished copy, just facts.
  • Resolve. If the issue is on our infrastructure, we fix it. If it's on an upstream provider we can't control, we say so and link to their status page.
  • Post-mortem. Any incident longer than 10 minutes gets written up in the list above — plain language, no spin, with what we changed to prevent repeats.

Need an SLA, procurement-grade uptime history, or a formal incident-response commitment? Email us and we'll tell you honestly whether we can meet your requirements today.