Anthropic enables Claude to control your Mac as research preview
Anthropic is rolling out computer control capabilities for Claude on macOS, allowing the AI to autonomously handle tasks like file navigation, clicking, and software interactions. The feature launches as a research preview for Claude Pro and Claude Max subscribers, with control available from iPhone via a new Dispatch tool.
Anthropic Enables Claude to Control Your Mac as Research Preview
Anthropric is introducing computer usage capabilities for Claude on macOS, allowing the AI assistant to autonomously control your Mac for task completion. The feature is rolling out as a research preview for Claude Pro and Claude Max subscribers.
The new functionality comes in two variants: Claude Cowork and Claude Code. Users can instruct Claude to perform tasks like opening files, editing documents, navigating applications, and handling complex software workflows. Control can be initiated directly from iPhone via a tool called Dispatch, which sends prompts to Claude for execution on the connected Mac.
Anthropric demonstrated the capability in an unlisted YouTube video showing practical computer usage scenarios. Each action requires initial user authorization before Claude executes it on the system.
Positioning in Competitive Landscape
This move places Anthropic directly in competition with OpenAI's Computer Use feature and other AI agents. Perplexity has introduced Perplexity Computer, while Meta is developing Manus for similar autonomous computer interaction. The AI agent space has accelerated significantly, with hardware constraints becoming a limiting factor—Mac minis, which power many of these systems, remain consistently sold out.
Anthropric's research preview approach differs from some competitors' immediate general availability, suggesting the company is iterating on safety and accuracy before broader rollout.
What This Means
Anthropric is extending Claude's capabilities beyond text generation into practical automation. Availability limited to paid tier subscribers ($20/month for Claude Pro, higher for Claude Max) positions computer usage as a premium feature. The iPhone-based control via Dispatch adds mobile accessibility to Mac automation, though the research preview status indicates reliability and safety features remain in development. This directly demonstrates that AI agents controlling personal devices are moving from theoretical to practical deployment, with the main constraints now being hardware availability and user trust rather than technical feasibility.
Related Articles
Anthropic Adds Cross-Session Messaging to Claude Code v2.1.224
Claude Code v2.1.224 introduces cross-session messaging, letting separate Claude Code instances on macOS and Linux send each other summaries to coordinate work. The feature does not support approving permissions or executing commands remotely.
Anthropic Cuts False Positives in Fable 5's Biology Filter by 85%, Keeps Virology and Toxicology Blocked
Anthropic has cut false positives in Fable 5's biology safety classifier by roughly 85%, letting users ask about lab results, symptoms, and medical questions without being rerouted to the weaker Opus 5 model. Dual-use topics like virology, toxicology, and molecular design remain restricted, with Anthropic citing the difficulty of containing biological threats once released.
Anthropic SDK v0.121.0 Adds Session Budgets, Mid-Conversation Tool Changes, and GitHub Skills Auto-Loading
Anthropic released version 0.121.0 of its Python SDK on August 7, 2026, introducing a new beta for mid-conversation tool changes, session budgets, an advisor tool, pinned inference location, and skills auto-loading from GitHub. The update also removes retired Claude Opus 4.1 models from the API.
Study: Humans Approve 1 in 3 Malicious AI Coding Agent Commands in Browser Game Test
A browser-based game simulating Claude Code-style permission requests found that human reviewers approved roughly one in three malicious commands across more than 40,000 game sessions. The findings, alongside Anthropic's own telemetry showing 93% approval rates for permission prompts, highlight growing concerns about approval fatigue in agentic AI coding workflows.
Comments
Loading...