Let Claude use your computer from the CLI
We're an AI research company that builds reliable, interpretable, and steerable AI systems. Our first product is Claude, an AI assistant for tasks at any scale.
Claude Code with computer use capability allows developers to execute routine tasks, build features, and handle Git workflows through natural language commands. The AI agent understands the entire codebase and can take actions across development tools, reducing manual work and accelerating development cycles.
Team and Enterprise plans include Claude Code with computer use for multiple users, enabling organizations to deploy AI agents at scale for engineering tasks, with role-based access controls, audit logs, and HIPAA-ready compliance for regulated environments.
Computer use in Claude Code enables the AI to examine relevant files, run the login flow, detect errors (e.g., '401 Unauthorized: Token expired'), identify root causes, apply fixes, and test solutions—all through automated interaction with the development environment.
Computer use in Claude Code can examine code files, identify performance improvements, suggest optimizations, and apply changes automatically. Evidence shows 25-100% productivity gains and ability to convert project requirements into technical specifications and system architecture designs.
Claude Code supports connectors to Jira, Confluence, Atlassian tools, Cloudflare Workers, and Intercom, allowing computer use capabilities to manage issues, access documentation, track sprints, and coordinate workflows without context-switching between applications.