Skip to content

DEVELOPER OPERATIONS

AI provider status

Current status and reported incidents for AI APIs and coding tools. Filter by service or follow the providers you use.

Vendor-reported, not independent probes. Feeds cached for 60 seconds.

Keep your integration running

Before switching providers

Match the affected component to your endpoint. A coding-app incident does not establish an API outage. Check your own errors, request IDs and latency too.

Test the fallback model’s tool calling, structured output, context window and data policy. A router using the same underlying provider may share the outage.

Compare routing and fallback options →

Protect the budget

Set a retry limit, exponential backoff with jitter, a request deadline and a fallback spending cap. Avoid automatically retrying tool actions that may have already completed.

Keep subscription capacity and metered API spend separate. A Max account’s usage allowance is not a guaranteed API token budget.

Estimate fallback API costs →

Coverage and methodology

OpenAI and Claude are connected through their public JSON status feeds. Components are classified by reviewed IDs; unmapped components remain in All components. Unknown statuses stay unknown. The watchlist is stored only in this browser.

History may be truncated by the publisher. We do not infer uptime percentages or restoration promises. No automatic failover or notifications are configured here; subscribe on the official provider pages.

Read the JSON status snapshot →

Other official status pages: DeepSeek ↗ · OpenRouter ↗. Their feeds are not connected yet.

Search API Tracker

Curated starting points