This is a live tracker of outages and incidents across the major AI tools. While our home page answers whether a specific tool is up right now, this page focuses on what has actually gone wrong lately: the most recent incidents across every service we monitor, newest first. Incidents confirmed by the provider link to the official report; the rest were detected by our own repeated checks.
Each incident is drawn from the providers’ own status records and our independent checks. You get the date, how long it lasted, and a link to the official report where one exists. Use it to confirm an outage you are seeing, to check whether a problem earlier today has been resolved, or to get a sense of which AI services have been least stable recently.
Recent AI incidents
Latest AI incidents
- Bolt.new [Dependency Check] Cloudflare Provider confirmed
- Cursor Investigating service degradation, IDE Provider confirmed
- Lovable Elevated error rates on chat requests Provider confirmed
- Bolt.new [Dependency Check] Name.com Provider confirmed
- MiniMax Elevated errors on Music Generation Provider confirmed
- Cursor Investigating Elevated Cursor Grok 4.5 Errors Provider confirmed
- Ideogram Complete outage on 3.1 and its variants Provider confirmed
- ChatGPT We’re seeing elevated latency, timeouts, and interrupted streaming across gpt 5.1 mini model and gpt 4.1 mini Provider confirmed
- Bolt.new [Dependency Check] Supabase Provider confirmed
- Cursor Investigating service degradation, Cloud Agents Provider confirmed
- Bolt.new [Dependency Check] Cloudflare Provider confirmed
- Kimi 搜索请求出现大量报错 Provider confirmed
- MiniMax Elevated errors on Large Language Models (LLM) Provider confirmed
- Claude Elevated errors on Claude Opus 5 and Haiku 4.5 Provider confirmed
- Cursor Investigating service degradation, IDE Provider confirmed
- Bolt.new [Dependency Check] Anthropic Provider confirmed
- Claude Elevated errors on Claude Opus 5 Provider confirmed
- Bolt.new [Dependency Check] GitHub Provider confirmed
- GitHub Copilot Incident with GraphQL API Requests Provider confirmed
- Bolt.new [Dependency Check] Anthropic Provider confirmed
How we track AI outages
We check each AI tool around the clock, about once a minute, combining each provider’s official status page (where they publish one) with our own reachability test from outside their network. When a provider reports a problem, or our own checks fail several times in a row (a single failed probe is never treated as an outage), we record an incident with a start time and, once service returns, a resolved time. We keep the daily uptime record indefinitely, so the history here grows more complete over time. For the full detail, see our methodology.
To check the current status of a single tool, open its page from the live status grid, where you will also find its uptime history, user outage reports, and troubleshooting steps for when it is not working.