Browser AutomationCommercial

Anthropic Computer Use

Claude's vision-based desktop and browser control

by anthropicApp Infrastructure
Open in Builder →Website ↗

About

Anthropic's vision-based agent capability that controls a desktop or browser by reading the screen and emitting mouse/keyboard actions. Reached 72.5% on OSWorld and is the most-discussed browser-automation API in 2026.

Choose Anthropic Computer Use when…

  • •You need an agent that can drive any GUI, not just a DOM
  • •You want first-party Claude vision-based control
  • •You are okay with vision-driven trade-offs vs DOM-driven reliability

Builder Slot

How does your AI navigate the web?Optional for most stacks

AI-powered browser control for agents that need to navigate, extract, fill forms, and interact with any website

Dev Tools
Optional
App Infra
Optional
Hybrid
Optional

Other tools in this slot:

Browser UseStagehandSkyvernFirecrawlBrowserbaseOpenAI Operator

Stack Genome Detection

AIchitect's Genome scanner detects Anthropic Computer Use in your project via these signals:

npm packages
@anthropic-ai/sdk
pip packages
anthropic
env vars
ANTHROPIC_API_KEY

Integrates with (1)

Anthropic APILLM Infrastructure

Anthropic Computer Use is a capability of the Claude API — vision-based mouse and keyboard control is exposed as tools on the standard Claude API.

→ Build vision-controlled agents directly from the Anthropic API without a separate runtime.

Compare →

Alternatives to consider (3)

OpenAI Operatorcompare →Browser Usecompare →Stagehandcompare →

Pricing

Pay-as-you-goStandard Claude API rates

In 1 stack

Browser AI / Web Agent Stack

Badge

Add to your GitHub README

Anthropic Computer Use on AIchitect[![Anthropic Computer Use](https://www.aichitect.dev/badge/tool/anthropic-computer-use)](https://www.aichitect.dev/tool/anthropic-computer-use)

Explore the full AI landscape

See how Anthropic Computer Use fits into the bigger picture — browse all 207 tools and their relationships.

Explore graph →