Anthropic Computer Use

https://docs.anthropic.com/en/docs/agents-and-tools/computer-use

Anthropic's API-surface capability that lets Claude control a virtual desktop — see screenshots, type, click. Launched 2024-10 as a public beta. Not a finished IDE itself; it's the substrate that other harnesses (and Anthropic's own Claude apps) use to operate a computer.

At a glance

Type
Claude desktop-control capability (Anthropic API)
Tier
T1
Created
2024-10
Latest release
not applicable — not OSS
License
not applicable — not OSS
GitHub
not applicable — no GitHub repo
Pricing
API per-token + per-screen-action
Funding
Anthropic (parent — $13.7B Sep-2025)

Taxonomy

storage
kv
retrieval
injection
persistence
session
update
agent-controlled
unit
turn
governance
opaque
conflict
stateless

When to use

Optimised for: Claude controlling a computer screen

Anti-fit: stateless capability — caller manages memory + safety

Pros & cons

Pros

First and most-discussed computer-control LLM; substrate for many higher-level agentic apps; benchmark numbers are public.

Cons

Stateless — every implementer rebuilds the memory + safety layer; per-action API cost is unpredictable.

Claims & capabilities

Anthropic API capability; first computer-control LLM at a major lab; SWE-bench / OSWorld benchmarks shipped with launch.

Technical surface

API surface
Anthropic Messages API with computer-use tool
Backend storage
stateless API
Deployment
API (Anthropic Cloud) — runs on caller's machine via local environment
Embedding model
not applicable — not a memory product
Multi-tenancy
API tenant
MCP
searched not found
A2A
searched not found
OpenTelemetry
searched not found

Compare Anthropic Computer Use with…

Similar systems

Other agent ides & coding harnesses in the catalog, ranked by inbound references.

  • Bedrock AgentCore (AWS) T2

    AWS Bedrock AgentCore — managed runtime to deploy + scale + observe agents. Multi-component: Runtime, Identity, Memory, Gateway, Browser, Code Interpreter, Observability. Distinct from Bedrock Agents (older fixed-config product); AgentCore is the 2025 framework-agnostic runtime for any agent (LangGraph, CrewAI, Strands, OpenAI Agents SDK).

  • Claude Code (Anthropic) T1

    Anthropic's official agentic coding CLI — runs Claude as a long-running coding agent in the developer's terminal. Memory layer = CLAUDE.md (project + user scoped), agent skills, MCP servers. Released GA October 2024; widely adopted as the reference harness for Claude-as-coder. Plan Mode and Compute Use surfaces extend it from coding to broader desktop automation.

  • GitHub Copilot (Agent Mode) T1

    The original AI coding assistant (Copilot in editor; 2021). 2025 evolved into a full agentic surface: chat, completions, Agent Mode (multi-step), and Copilot Workspace. Backed by Microsoft. Distinct row from Workspace — this row covers the everyday in-editor + agent mode product, Workspace covers the hosted spec-driven workflow.

  • Aider (harness) T2

    OSS terminal coding agent by Paul Gauthier — one of the original AI-pair-programmer CLIs (2023). Architect/Editor modes; repo-map for context; uses any LiteLLM-compatible model. Memory via CONVENTIONS.md and .aider.conf.yml — author-edited markdown. Distinct from the separate Aider memory-row in "Coding-agent memory"; this row characterises the harness itself.

  • Amazon Q Developer T1

    Amazon's developer-focused agent — formerly CodeWhisperer (renamed 2024-Q4). Available as IDE plugins (VS Code, JetBrains, Visual Studio, Eclipse), in AWS Console, and CLI. Strong AWS-tooling integration: Q can read CloudWatch, troubleshoot IAM, suggest AWS-SDK code. Companion to Kiro for spec-driven flows.

  • Bolt.new (StackBlitz) — harness T1

    Cross-listing of Bolt.new at the harness layer. StackBlitz's WebContainer technology runs full Node.js in-browser; Bolt prompts the user for an app spec, then writes/edits/runs/deploys it entirely in-browser. The Coding-agent-memory row remains canonical for the memory framing.

Related systems

References (1)

Row last verified 2026-05-14. Catalog data is CC-BY-4.0 — see how to read this.