Claude AI users ran into another outage on April 20, the second major disruption in three days, as problems hit the assistant, coding tools and mobile apps. The outage began at 14:53 UTC, and users were still reporting access problems after Anthropic said its API had recovered.
Downdetector data showed complaints peaking at 6,500 users reporting issues, with people flagging trouble reaching Claude Chat, Claude Code and the mobile app. By 8:23 pm IST, the service had already become a daily frustration for some users who depend on it for work.
Anthropic acknowledged a Partial Outage on its status page and said elevated errors were affecting Claude.ai, API and Claude Code. The company identified authentication issues as the main problem, and by 15:20 UTC said the API had fully recovered but the main interface still had access problems. Later, it said login success rates to Claude.ai, including via Claude Code, had stabilized and the issue was being fully resolved.
The repeated failures come as enterprise adoption is accelerating, which makes reliability more than a nuisance. Earlier in April, Claude experienced a similar disruption that affected conversational and coding tools for a substantial period, and this latest incident renewed concern that the service is having trouble keeping pace with demand.
Anthropic said the Claude API had fully recovered as of 8:01 PT, or 16:01 UTC, but the broader outage showed the gap between backend recovery and a service that users can actually reach. For people trying to use Claude Chat or Claude Code on April 20, that difference mattered more than the status update.