Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

Computer use

Computer use is the ability of an AI model to operate a computer the way a person does, by looking at screenshots and sending mouse clicks and keystrokes, so it can work with software that has no API.

Also called: computer-use agent, CUA

By Divya Sharma · Checked Nov 15, 2026 · How we review

Most agents act through APIs and tools built for them. Many programs only have a graphical interface. Computer use lets a model work those programs directly.

How it works

The application gives the model a computer tool and a screen, usually a virtual machine or container. The model asks for an action, such as clicking at a point, typing text or pressing a key. The application performs it and returns a new screenshot. The model looks at the result and decides the next action. This is the same agent loop as other tool use, with the screen as the feedback. It relies on a vision-language model that can read interfaces.

How it differs from a browser agent

A browser agent controls only a web browser, often through the page structure rather than pixels. Computer use can drive any desktop app, at the cost of being slower and less precise.

Example

An agent fills in an old desktop accounting program from a spreadsheet, clicking through the same screens a clerk would.

Risks and limits

Computer use is slow and makes mistakes on unfamiliar screens. It is also exposed to prompt injection: text on a web page or in an image can override the user's instructions. Anthropic recommends running it in a dedicated virtual machine or container with minimal privileges and keeping it away from sensitive data. Benchmarks such as OSWorld measure how well models do.

Related terms

Browser agent
A browser agent is an AI agent that controls a web browser to complete tasks, such as searching, clicking, filling in forms and reading pages, often running a headless browser with no visible window.
AI agent
An AI agent is a system in which a language model decides its own next steps: it calls tools, reads the results and repeats until the task is done, instead of following a fixed script.
OSWorld
OSWorld is a benchmark for computer-use agents that runs real operating systems such as Ubuntu, Windows and macOS and scores agents on 369 open-ended tasks across web and desktop apps, checked by scripts.
Vision-language model
A vision-language model (VLM) is a model that takes images and text as input and produces text, so it can describe pictures, answer questions about them, read documents and extract data from screenshots.
Prompt injection
Prompt injection is an attack where text fed to a language model, typed by a user or hidden in a web page, file or tool result, changes the model's behaviour in ways its developer did not intend.

Sources