AI Wiki tool
See active incidents first, then inspect component health and recent outage history across major AI model, app, and inference providers.
Live official snapshot
This is provider-reported status from public sources, not an independent probe of provider infrastructure or your account.
Auto-refresh is on and runs every 2 minutes while this tab is visible.
Needs attention
3 active notices
ChatGPT, API, and Codex
Partial System Degradation
API, models, and Playground
Partial System Degradation
Inference APIs and hosted models
Official status temporarily unavailable
The official page reports selected public endpoints and hosted-model monitors.
Component details were not returned. Use the official status page below.
No recent incidents were returned by this official feed.
Claude, API, and Claude Code
All Systems Operational
Gemini app
No active Gemini incidents in the official Workspace feed
Google's public Workspace feed covers the Gemini app. Check Google AI Studio separately for Gemini API-specific notices.
Search, API, and Computer
All monitored Perplexity services operational
Perplexity's public API exposes active incidents and components, but not historical notices.
No active incidents were returned by this official feed. Historical notices remain available on the official page.
Hub, inference, Spaces, and Jobs
All monitored Hugging Face services operational
xAI API and Grok
No unresolved incidents in xAI's official feed
Synthesized from xAI's incident feed; xAI does not expose machine-readable component states. Its iOS and Android services are grouped as Grok mobile apps.
GroqCloud API and models
All Systems Operational
Inference, Playground, and models
All monitored Together AI services operational
Together AI publishes platform monitors plus model-specific availability checks.
API, inference, training, and models
All Systems Operational
Mistral AI API and Console
No unresolved incidents or active maintenance reported by Mistral
The component list focuses on eight core API and Console services. Individual states are synthesized from active incidents; otherwise the feed does not report a live component state.
OpenRouter API
All OpenRouter uptime checks passing
OnlineOrNot uptime checks on OpenRouter's public endpoints, not internal telemetry.
Chat and API
Not polled by this dashboard
DeepSeek's official status page is served from mainland China and is not reliably reachable from our servers; check it directly.
AI Wiki reads public status data published by providers or their named status-page vendors. It does not send test requests to provider infrastructure and cannot inspect your account, quota, region, or network.
What does Operational mean on this dashboard?
Operational means the provider publishes a normal component state. Incident-only and external-monitor feeds use more specific labels such as No active incident or Checks passing.
Where do the Degraded and outage labels come from?
Degraded and outage labels are normalized from provider-published component states, incident severity, or maintenance notices. Each card identifies important coverage limits.
What does "Not reported" mean?
"Not reported" means a feed did not expose a live state or could not be reached.
Can a provider look operational here while my requests fail?
Yes. Account-specific limits and regional failures may not appear on a global status page.
Which AI services are monitored?
The dashboard monitors public status sources for 13 providers: OpenAI, Anthropic, Google Gemini, Perplexity, Hugging Face, xAI, Cohere, Groq, Together AI, Replicate, Fireworks AI, Mistral, and OpenRouter. DeepSeek is linked directly but not polled.
Why is there no live DeepSeek status?
DeepSeek's official status page is served from mainland China and is not reliably reachable from our servers, so this dashboard links to it instead of polling it.