Skip to main content
Live
Main content

Anthropic frees Claude Cowork from the desktop with always-on mobile agent

Cowork now runs tasks overnight without an open laptop, arriving first to Max plan subscribers at $100 a month.

Jaeden Schafer
Editor in Chief · · 4 min read
Anthropic logo

Anthropic on Tuesday cut the desktop tether on Claude Cowork, its agent for doing digital work on a user's behalf. The agent can now run scheduled tasks without a laptop kept awake and open, and Anthropic is shipping trimmed-down versions of Cowork inside the existing Claude mobile app and browser. The beta lands first on the Max plan at $100 a month, then trickles down to Pro subscribers at $20 a month.

The pitch is straightforward: users prompt Cowork from a phone, close the phone, and come back to a finished job. Anthropic's demo shows a single request pulling data from email threads, Slack channels, meeting transcripts and recent online chatter, then producing a briefing document and a prewritten email for a next-day business renewal. Previously that workflow required an active desktop session, which is why some users left laptops cracked open overnight.

Cowork first shipped in January 2026 and could already complete real desktop tasks — organizing screenshot files into labeled folders, populating a calendar — as long as the machine stayed on. Mobile access existed through a pairing feature called Dispatch, which forwarded task requests from phone to desktop.

Key facts

  • 01Claude Cowork can now run scheduled tasks without an active desktop session, ending the laptop-cracked-open workaround.
  • 02The revamped Cowork rolls out first as a beta to Max plan subscribers at $100 a month, then to Pro at $20 a month.
  • 03Anthropic says business process work and content creation are the two largest usage categories for Cowork.
  • 04OpenAI shipped Codex Remote last month and added Codex task management to its iOS app in July.
  • 05Google launched Spark as its always-on agent entry earlier in the first half of 2026.

The catch was spelled out in Anthropic's own product documentation for Dispatch.

Your computer must be awake, and the app must be open for Claude to work on tasks
Anthropic, product documentation for the prior Dispatch feature

Removing the awake-desktop requirement is what makes today's release material. Cowork can now catch late-night incoming messages and execute overnight without babysitting hardware, which is the difference between a demo and a workflow people actually build around.

The competitive frame is clear. OpenClaw, a homebrew agent that went viral at the beginning of 2026, established the always-on, text-controlled agent as the template early adopters wanted. OpenAI responded by hiring OpenClaw's creator and launching Codex, then rolled out Codex Remote last month to let users control desktop agents from a phone. In July, OpenAI added support for creating, searching, opening, forking and managing Codex tasks directly from a conversation inside its iOS app. Google launched Spark as its own always-on agent during the first half of the year.

Anthropic's own path into this space ran through Claude Code, its developer-focused automation tool. Cowork applies the same execution loop to non-developers, swapping the terminal for a chatbot interface and targeting knowledge workers instead of engineers.

The company also released a usage report alongside the launch. Anthropic said business process and operations work — data reports and checklists — and content creation and copywriting — slide decks and partnership proposals — are the two largest categories of recent Cowork usage. That maps to the white-collar workflows Anthropic has been steering the product toward.

Related · from this week
Anthropic's Claude agents started a turf war when set loose on the same task
Jaeden Schafer · 5 min read →

There are real limits worth naming. The agent still exposes users to prompt-injection risk and other security surfaces that come with any tool that reads email, Slack and web content on a user's behalf. Anthropic has not said whether free-tier users will ever see Cowork; today they don't have access, and the rollout order — Max first, then Pro — suggests the compute cost of running unattended agents is still high enough to gate behind the top plan.

The strategic shift is that Anthropic and OpenAI are both racing to fold agents into the chat apps hundreds of millions of people already open every day, rather than shipping them as separate developer tools. Whichever company gets always-on mobile agents to feel reliable first captures the default surface for a huge chunk of everyday knowledge work — and Anthropic's decision to merge the Claude chatbot and the Cowork agent inside a single interface is a bet that consolidation, not proliferation, is what turns agents into a habit.

ShareXLinkedInEmail
AI Box

Every AI model. One chat.

The latest models from ChatGPT, Claude, Gemini, Sora, ElevenLabs — 80+ models in a single chat. Compare answers side by side. Pick the best one every time.

  • ChatGPT, Claude, Gemini, Grok, DeepSeek — in one chat
  • Generate images & video with Sora, Veo, Ideogram
  • Compare any two models side by side
  • From $8.99/mo · 80+ models, all included
Try AI Boxaibox.ai
Trusted by 3,000+ teams
Got a tip?

Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.

Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.

AI Box Daily briefingFree · Daily · No fluff

Stay ahead of everyone in AI.

The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.

Loved by 10,000+ AI professionals
Free forever. Unsubscribe with one click.

The briefing read inside teams at

Keep reading

More from Models

Anthropic logo
Security

Anthropic's Claude agents started a turf war when set loose on the same task

Frontier Red Team found agents with conflicting instructions sabotaged each other with self-replicating malware — and sometimes negotiated truces.

Jaeden Schafer5 min read
Anthropic logo
Models

Anthropic merges Claude and Claude Cowork memory into one system

Claude will now carry context across chat and Cowork, and users can read, edit, or delete stored memories on any topic.

Jaeden Schafer4 min read
OpenAI logo
Models

OpenAI ships GPT-5.6 in three tiers, undercuts Claude on price

Sol, Terra, and Luna launch under a White House-monitored preview, with Sol priced at half of Anthropic's Claude Fable 5.

Jaeden Schafer5 min read