AI TraceTrace Foundation, Inc.
Productivity AutomationReplaces Human LaborVerified

Reviewed and published by trentmaziarz, April 8, 2026. Discovered and drafted by our automated research pipeline.

Anthropic offers Claude Cowork, a desktop agent for knowledge workers that reads, edits, and creates files, navigates applications, and completes multi-step tasks without the user managing each step. It is available to paid Claude subscribers (Pro, Max, Team, and Enterprise) through the Claude Desktop app on macOS and Windows, launched as a research preview in January 2026. The underlying Computer Use capability, which lets Claude perceive a screen and control a keyboard and mouse, launched as an API beta in October 2024.

Details

Anthropic's Computer Use feature launched as an API-only beta in October 2024, making Claude 3.5 Sonnet the first frontier AI model to offer computer use in public beta. Claude Cowork followed on January 12, 2026 as a tab inside the Claude Desktop app, aimed at knowledge workers rather than software developers; it launched initially for Max subscribers on macOS and expanded to all paid plans on both macOS and Windows by April 2026. On the OSWorld benchmark, which measures an AI's ability to complete real tasks on a computer such as editing documents and navigating browsers, Claude Sonnet 4.6 scored 72.5%, matching the human baseline of approximately 72% on that evaluation, up from 14.9% when computer use launched in October 2024. Anthropic acquired Vercept, a Seattle startup focused on vision-based computer automation, on February 25, 2026 to advance these capabilities. Following the January 30, 2026 release of industry-specific Cowork plugins, shares of enterprise software companies including Thomson Reuters, RELX, and LegalZoom fell sharply, as investors anticipated AI competition with knowledge-work software.

Products affected

Claude CoworkClaude Desktop AppClaude Computer Use APIClaude for Chrome

Sources & Evidence

News Article

Company Disclosure

Other practices by Anthropic

Have evidence about Anthropic's AI practices? Submit a report.

Report a Sighting →