"API Error: 529 Overloaded. This is a server-side issue, usually temporary — try again in a moment," users reported as Anthropic confirmed elevated errors across multiple AI models on July 29. The message — and its implications — cut off access to Claude and to tools that depend on Claude’s API, leaving developers and end users waiting for a fix.
Anthropic’s timeline of the outage
Anthropic began investigating the disruption at 7:49 p.m. UTC on July 29, according to the company’s status updates. By 8:33 p.m. UTC the company said it had identified the issue and was working to resolve it, but did not disclose the underlying cause or offer a recovery timeline. Those two timestamps are the only firm chronological details the company provided in its public notices.
Symptoms: 529 Overloaded and failed requests
The visible symptom reported across affected services is a consistent failure of requests accompanied by the message quoted above. The outage causes requests to fail with a “529 Overloaded” message; the source explains that "529" indicates Claude’s servers are unable to handle the current volume of requests. For affected users, refreshing the page does not clear the error or restore service.
Scope: Claude, API tools, and all Anthropic-hosted services
Reports indicate the problem is not limited to an individual model instance. The disruption is causing failures "including in Claude and tools that rely on its API," and the outage "appears to affect all services hosted by Anthropic." That phrasing signals a platform-wide impact rather than a narrowly contained server fault.
Anthropic’s public response and remaining unknowns
Anthropic acknowledged elevated errors and confirmed it was investigating. The company said it had "identified the issue and was working to resolve it" but did not reveal the technical root cause or provide an estimated recovery time. Separately, the report includes the statement "Thankfully, Anthropic is aware of the root cause and is working on a fix," echoing the company’s message of active remediation while underscoring that details remain unpublished.
What this means for Claude users, API integrators, and enterprises
- Claude users: Individuals relying on interactive sessions with Claude experienced immediate interruptions — requests failed with the 529 message and page refreshes did not restore access.
- API integrators and tool builders: Tools that depend on Claude’s API likewise failed to complete requests, meaning downstream applications and integrations were disrupted while the outage persisted.
- Enterprises using Anthropic-hosted services: Because the disruption "appears to affect all services hosted by Anthropic," organizations that route workflows through Anthropic’s platform faced broader service interruption until the provider restores capacity.
The outage timeline is narrow but clear: investigation began at 7:49 p.m. UTC and Anthropic reported identification of the issue by 8:33 p.m. UTC on July 29. What the company has not provided is a technical explanation or a recovery schedule — leaving users and integrators with the immediate fact of service interruption and the single reassurance that Anthropic was working to fix the problem.
For now, the concrete questions left by the record are simple and specific: when will Anthropic restore normal capacity across its hosted services, and what operational or engineering change will prevent a recurrence? Until the company publishes those details, the outage stands as a live disruption with known timing and symptoms but without a disclosed cause or timeline for recovery.
Original reporting: https://www.bleepingcomputer.com/news/artificial-intelligence/anthropic-confirms-claude-is-down-worldwide/
