Live health,
checked in real time.
This page queries the platform directly every sixty seconds. Nothing here is hand-maintained, so it cannot quietly go stale.
Current status
—
systems monitored
—
fully operational
Checking…
Components
Each system, individually.
Database
Accounts, generations, credit ledger
Cache & rate limiting
Queue state, rate limits, prompt cache
Image generation
Text-to-image and editing models
Video generation
Text-to-video and image-to-video models
Content moderation
Prompt safety screening
How to read this page
- Operational — the system responded to a live health check within its timeout.
- Degraded — fallback active — a component is unavailable and the platform is running on its backup path. Things still work, sometimes more slowly.
- Degraded — the check failed. If this is the database or a generation service, expect errors.
Individual model availability is separate from this page. Models are health-probed continuously and anything failing upstream is hidden from the gallery automatically, so a single misbehaving model does not show up as a platform outage.
If something is broken for you
A render that fails always refunds its credits automatically — you are never charged for output you did not receive. If a job has failed and the credits have not come back within a minute, that is a bug worth reporting rather than expected behaviour.
The fastest route is the in-app assistant, which can see your account context. For anything that looks like a platform-wide problem, use the contact form and include the generation id.
Planned work
Deploys are rolling and do not require downtime. Anything that will interrupt service is announced in the changelog ahead of time.