Anthropic has released Claude for Windows, expanding earlier Mac availability, enabling Claude Cowork and Claude Code to automate tasks by opening apps, debugging code, and handling other hands-free actions.

The Windows feature is available to Claude Pro and Claude Max subscribers after downloading the app, with no extra setup beyond installation.

Dispatch lets users assign tasks remotely via smartphones, enabling productivity across devices.

Anthropic addressed user concerns about faster-than-expected token usage during peak hours, confirming no extra charges and explaining adjustments like tighter peak-hour limits and larger 1M-context sessions.

Key concerns for broader adoption include data privacy, potential error rates, and controllability, indicating the technology remains unproven for widespread real-world use beyond the research preview.

Anthropic highlighted culture and talent as a competitive advantage, with Vercept’s co-founder emphasizing speed and collaboration within the company.

Industry reflections compare this to OpenAI's browser-based agents and emphasize the significant attack surface and reliability challenges of full desktop control.

The information comes from Anthropic's official X announcement and subsequent clarification from Lydia Hallie.

There are reports that AI usage quotas are being consumed more quickly than normal, and Anthropic is investigating these issues.

The Windows rollout follows the Mac release at the end of March, with ongoing user feedback and refinements, including rate-limit considerations and model recommendations for power users.

For security, users must approve app access at the start of each session, with per-session permissions and the ability to abort actions via Esc or terminal commands; only one Claude session can control the computer at a time.

Operations are restricted to a single active Claude instance to prevent conflicts, with safeguards like per-session approvals and immediate halting options.