AI developer tools status
Is it down, or is it you? One board for the tools you actually ship with — component-level, straight from each vendor's official status API.
Your browser reads the official APIs directly. Nothing passes through our servers, and we never guess a status.
Nearest official signal: Gemini API on Google Cloud
How this board works
Every tile reads a vendor's official status API directly in your browser — status.claude.com, status.openai.com, status.cursor.com, githubstatus.com, status.windsurf.com, and Google Cloud's incident feed for Antigravity. We don't run our own probes, and nothing passes through our servers, so what you see is the same signal you'd get opening each page yourself, just collected in one place.
The board refreshes automatically every 60 seconds while the tab is open. If a fetch fails, a tile reads "Unreachable." and links straight to the official page instead of guessing. Component filters — Copilot on GitHub's page, Claude Code on Anthropic's — mean each tile reflects the piece you depend on, not just a headline "all systems operational" banner.
Is it down or is it just you?
Before blaming a vendor, rule out the usual local suspects: an expired auth token, a stale extension build, a hit rate limit, or a flaky network. Those produce the same errors as a real outage but won't show up on any status page, official or otherwise.
If the board shows every component green and your tool still won't respond, the problem is almost certainly local — try re-authenticating, checking your usage quota, or restarting the client. If the board shows a live incident matching your symptoms, the practical fix is usually to wait it out, switch models where the tool allows it, or fall back to another tool for anything urgent. Either way, the official status page linked on every tile is the final word, not us.
Frequently asked questions
How does this status board work?
Your browser reads each vendor's official status API directly — status.claude.com, status.openai.com, status.cursor.com, githubstatus.com, status.windsurf.com, and Google Cloud's incident feed. Nothing passes through our servers, and when an API can't be reached we say so instead of guessing.
Why component-level status?
Because "is Claude down" is usually the wrong question. Claude Code can be degraded while claude.ai is fine, and GitHub can be green while the Copilot component isn't. Component rows answer what you actually ship with.
Is this an uptime monitor?
No. We don't run probes or collect user reports — the board mirrors official vendor status feeds in real time and computes 30-day incident stats from their published history. For alerting, subscribe on the vendor's own status page.