Anthropic's Claude AI experienced a significant outage on April 15, 2026, affecting Claude.ai, its API, and Claude Code services. Thousands of users reported access issues, with Downdetector showing a sharp spike in complaints during the morning (PT) hours.
Early detection and investigation
The incident was first flagged at 14:53 UTC, when increased error rates were observed across all major Claude services. By 14:55 UTC, the company confirmed it was investigating the issue.
Rapid escalation of user complaints
User reports surged quickly as the outage expanded:
7,000+ reports by 8:02 a.m. PT
9,000+ reports by 8:08 a.m. PT
15,000+ reports by 8:25 a.m. PT
20,000+ reports by 8:39 a.m. PT
The spike indicated a widespread disruption impacting users globally.
Services affected
Users experienced multiple issues across the platform, including:
-Connectivity failures on Claude.ai
-Login issues affecting Claude Code and web access
-General website accessibility problems
-Partial recovery and API restoration
By 15:20 UTC, the issue had been identified and a fix was underway. At 15:40 UTC, Anthropic confirmed that the Claude API had fully recovered as of 16:01 UTC (8:01 PT). However, login issues and platform instability continued for Claude.ai and Claude Code.
Continued instability and recovery progress
At 16:29 UTC, Claude.ai and platform access were still down, though logged-in Claude Code users could continue working. Engineers reported improving login success rates shortly after as systems stabilized.
Full resolution
At 17:42 UTC, the incident was officially resolved. All services, including Claude.ai, API, and Claude Code, were restored following several hours of intermittent disruption.