Anthropic brings Claude Cowork agent to mobile and web with background task execution
Anthropic is expanding Claude Cowork from desktop-only to mobile and web platforms. The autonomous agent feature will run tasks in the background across devices, with scheduled execution requiring no device to be online, starting with Max plan subscribers in a gradual rollout over the next several weeks.
Anthropic Brings Claude Cowork to Mobile and Web with Background Execution
Anthropic announced that Claude Cowork, its autonomous agent feature previously limited to the desktop app, is expanding to web browsers and mobile devices with new background execution capabilities.
The feature enters beta with a gradual rollout starting with Max plan subscribers over the next several weeks, according to the company's July 7 announcement. Additional paid plans will gain access in subsequent phases.
Core Capabilities
According to Anthropic, Claude Cowork "works across your files, calendar, email, messaging app, the web, and the other tools you connect until the job is done." The mobile and web expansion introduces three technical changes:
Cross-device continuity: Tasks initiated on desktop can be monitored from mobile devices, with outputs accessible across all platforms through claude.ai and native iOS, Android, and desktop apps.
Background execution: Tasks continue running when devices are closed or disconnected. The system no longer requires an active client connection to complete work.
Scheduled execution without online devices: Users can schedule tasks to run at specific times with no device required online. Anthropic's example: "Set Monday's client prep for 6 am: Claude works through the email threads, transcripts, and recent news, builds the briefing doc, and leaves the follow-up email drafted but unsent."
The system maintains human approval gates. When decisions are required, Cowork sends mobile notifications. No outputs are sent or published without explicit user review and approval.
Interface Changes
Mike Krieger, Anthropic product executive, stated that Chat and Cowork now share a unified interface on web and desktop platforms, consolidating the sidebar, search functionality, and Projects & Artifacts into a single tab. Further Chat and Cowork integration is planned.
Cowork will appear in the sidebar of the Claude mobile apps on iOS, iPadOS, and Android, alongside web access at claude.ai.
Availability
Anthropic is extending doubled Cowork usage limits through August 5 to mark the launch. Pricing per task and specific usage quotas were not disclosed in the announcement.
What This Means
The shift from desktop-only to cloud-based execution fundamentally changes Claude Cowork's architecture from a locally-tethered agent to a server-side autonomous system. Background execution removes the constraint of maintaining an active client session, making scheduled workflows practical for enterprise users. However, the requirement for explicit approval before any external action limits full autonomy, positioning Cowork as a supervised agent rather than a fully autonomous one. The staggered rollout by plan tier suggests Anthropic is managing infrastructure load as it scales server-side agent execution.
Related Articles
Anthropic Report: Claude Was Used to Target US Navy Ships, Build Missiles, and Track Uyghurs
Anthropic's latest threat intelligence report documents five cases where state and non-state actors used Claude for military targeting, weapons development, mass surveillance, and repression. The findings include an Iran-linked operation targeting US naval forces and a Mali-based system capable of monitoring 25 million phones.
Anthropic Threat Report: Claude Used for Missile Software, Mass Surveillance, and Systematic Theft by Chinese AI Labs
Anthropic's latest threat intelligence report covers December 2025 through August 2026, documenting Claude's misuse in espionage, weapons development, and nationwide surveillance operations. The report also details how seven Chinese AI labs ran covert networks—some routing their own customers' requests through Claude—to extract training data at industrial scale.
Analysis: Claude 'Fable 5.1' Drops Em Dashes and Hedging Language, Answers Grow 30% Longer
A new Arena.ai analysis of tens of thousands of Text Arena outputs shows Claude 'Fable 5.1' has shifted its writing style significantly from Fable 5 — using fewer em dashes, less hedging language, and producing 30% longer responses. The codenamed models appear to be unreleased Anthropic checkpoints being tested anonymously on LMArena.
Anthropic CEO Dario Amodei Proposes Three-Step Plan to Deliberately Slow AI Capability Advances
Anthropic CEO Dario Amodei published an essay proposing a three-step plan to deliberately pace AI development, including third-party safety audits and cross-industry coordination. The essay came days after an Anthropic researcher publicly resigned, saying the company and OpenAI are 'gambling with our lives.'
Comments
Loading...