Key Highlights
- Anthropic says Claude can now use your computer to complete tasks autonomously after receiving prompts from a smartphone.
- The feature allows Claude to open apps, browse the web, and edit files locally on a user’s device.
- The update strengthens Anthropic’s push into AI agents, following the viral rise of OpenClaw.
- Anthropic warns the capability is still early and includes safeguards to reduce risk.
Anthropic says Claude can now use your computer to finish tasks on your behalf, a move that pushes artificial intelligence closer to fully autonomous digital agents. The company announced that users can send prompts from their smartphones, while Claude carries out the work directly on their personal computers.
The update, revealed Monday, positions Claude as a hands-on AI agent capable of interacting with apps, browsers, and files locally, without constant user supervision.
What exactly did Anthropic announce?
Anthropic says Claude has gained a new “computer use” capability. Once a user assigns a task from a phone or desktop, Claude can operate the user’s computer to complete it.
According to the company, Claude can:
- Open and navigate desktop applications
- Use a web browser to find information
- Edit documents and fill spreadsheets
- Export files and attach them to emails or calendar invites
In one demonstration video, Anthropic showed a user running late for a meeting. The user asked Claude to export a pitch deck as a PDF and attach it to a meeting invite. Claude completed the entire task without further input.
This marks a shift from chat-based assistance to action-based execution.
Why is “Anthropic” suddenly trending?
“Anthropic” is trending as competition intensifies around AI agents. These are systems designed to perform multi-step tasks independently, rather than just respond to prompts.
The concept gained widespread attention earlier this year after the viral release of OpenClaw. OpenClaw allows users to message an AI agent through apps like WhatsApp or Telegram and have it perform tasks locally on their devices.
Anthropic’s latest Claude update directly targets this emerging category.
By allowing Claude to control a user’s computer, Anthropic signals that it wants to be a central player in the race to build always-on, task-completing AI agents.
How does Claude’s computer use feature work?
Anthropic says users can message Claude a task from their smartphone using a continuous conversation feature. Claude then executes the task on the user’s computer.
The system runs locally, meaning Claude has access to files and applications already on the device. This mirrors how OpenClaw operates and reduces reliance on cloud-only workflows.
Anthropic also introduced a tool called Dispatch, available through Claude Cowork. Dispatch lets users maintain an ongoing conversation with Claude across devices and assign tasks at any time.
This approach positions Claude less as a chatbot and more as a digital operator.
How does this compare with OpenClaw and rivals?
OpenClaw has become a reference point in the AI agent conversation. It connects to models from OpenAI and Anthropic and executes tasks locally on a user’s device.
Nvidia CEO Jensen Huang recently called OpenClaw “definitely the next ChatGPT,” underscoring how seriously the industry views agent-based AI.
Nvidia has since announced NemoClaw, an enterprise-grade version aimed at business users. Meanwhile, OpenAI hired OpenClaw creator Peter Steinberger last month to accelerate its own personal agent strategy.
Anthropic’s Claude now joins this competitive field with native computer control, signaling that AI agents are no longer experimental side projects.
What safeguards does Anthropic say are in place?
Anthropic emphasized that computer use is still an early capability.
The company warned that Claude can make mistakes and that threats continue to evolve as AI becomes more capable. To address this, Anthropic says it has built safeguards to minimize risk.
Key safeguards include:
- Claude requesting permission before accessing new apps
- Limits on sensitive actions
- Ongoing monitoring and improvement of safety systems
Anthropic noted that Claude’s computer use is less mature than its coding or text-based skills.
The company framed the rollout as a trial rather than a finished product.
Why does this matter for everyday users?
For users, the update hints at a future where AI handles routine digital tasks in the background.
Instead of manually opening apps, exporting files, or organizing meetings, users can delegate those steps to an AI agent. The promise is saved time and reduced friction, especially for repetitive work.
At the same time, Anthropic’s caution highlights the risks of giving AI deeper system access. Mistakes, misinterpretation of instructions, or security concerns remain unresolved challenges.
What does this mean for the AI agent race?
Anthropic says Claude’s new capability reflects a broader industry shift toward autonomy. AI companies are racing to build agents that can work continuously, across devices, and with minimal oversight.
This shift moves AI from passive assistant to active participant in digital workflows.
As Anthropic continues testing Claude’s computer use feature, the focus will likely remain on reliability, safety, and user trust.
In the rapidly evolving agent landscape, Anthropic says Claude is no longer just answering questions. It is starting to act.
