Detailed Analysis
A recurring user experience issue has emerged around Anthropic's Claude Desktop application, specifically concerning the visibility and accessibility of the "Computer use" feature for newly subscribed Pro plan users. The feature, which enables Claude to directly interact with a user's computer — performing tasks like navigating files, opening applications, and executing commands — is not enabled by default and does not surface automatically in the Desktop app's settings upon upgrading to a paid tier. Users who upgrade expecting immediate access are instead met with a settings panel that appears to lack the feature entirely, creating confusion about whether the capability is included in their plan or whether a technical fault is responsible.
The root cause in most cases is a combination of outdated app versions and the feature's opt-in design. Anthropic has structured Computer use as a toggle under Settings > General within the Desktop app, but that toggle only appears when the application is fully up to date and has been restarted after the subscription upgrade is recognized. The feature is restricted to Pro, Max, Team, and Enterprise plan holders on macOS and Windows, meaning free-tier users will never see the option regardless of app version. For users on Windows in particular, installation integrity issues add a compounding layer of complexity — a failed or incomplete MSIX package installation can prevent critical app components from rendering correctly, requiring administrative reinstallation or PowerShell-based package removal before a clean install can succeed.
The troubleshooting pathway Anthropic recommends follows a logical escalation: first, confirm the paid plan is active and the app is updated via claude.com/download; second, fully quit and relaunch the application; and third, navigate to the settings toggle directly. For Windows users encountering deeper issues, enabling Developer Mode or using PowerShell to strip out corrupted package remnants before reinstalling represents a more advanced remediation step. The fact that these workarounds are circulating in community forums and GitHub issue threads as of early 2026 suggests that the onboarding experience for Computer use has not yet been streamlined to match user expectations upon plan upgrade.
This pattern reflects a broader challenge Anthropic faces as it scales agentic capabilities — features like Computer use represent a significant leap in AI autonomy and carry meaningful risk implications, which justifies a conservative opt-in design. However, that same conservatism creates friction for users who have paid for access and cannot easily discover or activate the feature. The gap between a user's expectation upon upgrading ("I should now be able to use Computer use") and the actual multi-step activation process points to an onboarding and discoverability problem that is distinct from the technical issues themselves. As agentic AI tools become more central to Anthropic's product offerings, the user experience around feature discovery and activation will likely need to evolve alongside the capabilities themselves.
Read original article →