Claude Code outage: users complain about access problems
The Claude Code service has temporarily stopped working: some users cannot log in to their accounts or use the tool.
The Claude Code service has temporarily stopped working: some users cannot log in to their accounts or use the tool.
The Claude Code service has temporarily stopped working: some users cannot log in to their accounts or use the tool.
According to Anthropic’s status page, the outage affected several of the company’s products, including Claude.ai, API, and Claude Code. The company confirmed the incident and reported an increased number of errors in the services.
Users encounter typical problems: the system does not allow them to log in to their account, pages do not load or freeze after trying to log in. In some cases, the service opens, but works unstable and does not respond to actions.
On social media and forums, including Reddit, users are reporting massive outages. They report endless login attempts, connection errors, and situations where the service simply doesn’t respond. Some developers say that even when they have access, the tool doesn’t respond or is slow to respond.
Anthropic said it is working to resolve the issue. Some functionality has been restored, but access disruptions for some users remain.
Similar problems with Claude’s work occur regularly in 2026. In April alone, several failures related to access to the service were recorded: on April 13, individual users could not log in to their accounts, and earlier there were problems with the API and web version. In March, there were also large-scale outages, when the services were partially or completely unavailable for several hours.
Recent incidents have a similar pattern: authorization issues, connection errors, and unstable operation under load. This indicates that the failures are not isolated and are recurring with varying intensity.
Previously, dev.ua wrote about how an error during the publication of the package led to a leak of the internal code of the Claude Code tool, which allowed potential attackers to study its operation from the inside.



