Computer Use
Computer Use is an approach in which an AI agent reads a screen and operates a computer through mouse and keyboard actions.
Computer Use is an approach in which an AI agent interprets screen images and generates mouse and keyboard actions to operate apps or websites. The term gained wide currency after Anthropic introduced a feature by that name with Claude 3.5 Sonnet in October 2024. Rather than calling a service's API directly, the agent works through the same graphical interface a person would use. An AI agent that works this way is called a computer-use agent.
Unlike conventional screen automation that replays predefined actions, the model reads the changing screen and chooses its next step. Performance is compared on benchmarks such as OSWorld, which test task completion in real computer environments.
This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.
Articles covering this entry
Computer Use API: an API for computer use, where an AI model operates a computer directly
Turn off the embedded browser under the "Computer Use" section of the settings.
OpenAI's internal developer experience team already relies on computer use agents to run its YouTube operations.
The Agents API adds Computer Use, which lets an agent interact with a graphical interface when an application lacks a suitable developer API.
Chrome Computer Use is a separate, broader permission decision.
GPT-6 Astra’s computer-use results make it worth assessing for repetitive work, but the Kit test shows why that assessment should include what the agent is allo…
The reported comparison is specific to the computer-use test, not a measured speedup for every Jev application.
Yutori’s Navigator n2 shows why computer-use agents need to be judged on cost and reliability together.