Everything you need to know about Claude Cowork, including features, pricing, and how to access the new productivity tool
The new Claude Cowork tool can work autonomously on basic tasks like sorting folders
Sign up today and you will receive a free copy of our Future Focus 2025 report - the leading guidance on AI, cybersecurity and other IT challenges as per 700+ senior executives
You are now subscribed
Your newsletter sign-up was successful
Anthropic wants users to hand Claude their files and let the AI bot figure out what to do with them – but it warns that the system might choose to delete your files.
The AI developer this week unveiled Claude Cowork, which applies abilities created for Claude Code to non-coding tasks in an "approachable" way, the company said in a blog post, once users have given access to files, the wider web, or third-party data.
Tasks could include generating reports, organizing files, or creating a spreadsheet based on documents.
Anthropic said Claude Cowork was inspired by the way developers used Claude Code, the company's AI code generation tool.
"When we released Claude Code, we expected developers to use it for coding," Anthropic said. "They did – and then quickly began using it for almost everything else."
The aim, according to Anthropic, is to make it easier to put AI to work without constantly giving more context or instructions, or having to convert outputs into documents yourself.
"It feels much less like a back-and-forth and much more like leaving messages for a coworker," the company said.
Sign up today and you will receive a free copy of our Future Focus 2025 report - the leading guidance on AI, cybersecurity and other IT challenges as per 700+ senior executives
What can Claude Cowork do?
Claude Cowork isn't just inspired by Claude Code, but built on the same "foundations", the company said – and that means the system will feel familiar to anyone who has already used the coding tool.
Notably, Cowork differs from the standard chatbot as it works by giving it access to a specific folder on your computer, in order to read and edit files – as well as create its own, as Anthropic has built-in skills for making presentations and other documents.
"It can, for example, re-organize your downloads by sorting and renaming each file, create a new spreadsheet with a list of expenses from a pile of screenshots, or produce a first draft of a report from your scattered notes," the company said.
Anthropic said its Claude AI will then work with "more agency" than a chatbot, and multiple tasks can be queued up at once for the system to do at the same time.
"Once you’ve set it a task, Claude will make a plan and steadily complete it, while looping you in on what it’s up to," the company said.
Beyond desktop files, Cowork can access external files by using existing connectors, and can be paired with Chrome for web access.
How to access Claude Cowork
Claude Cowork remains in research preview and is available to Claude Max subscribers.
However, the company noted that access will be limited to those on the macOS app to begin with. Anthropic said it plans to offer access to Windows users too, but exact details on a release timeline are yet to be determined.
The aim here for Anthropic is to get a sense of how people use Cowork to drive further improvements and development before a full release.
Anthropic's warning
Claude Cowork will only access files that it's been given direct access to view, and will ask for approval at key stages of a task.
Anthropic did warn that users should perhaps limit the tool's use to a single folder, and urged users to keep separate backups of the files it's been set to work on. The reasoning here is that the bot could take "potentially destructive actions" such as deleting materials.
The company said Claude would only delete a file if instructed, but added: "Since there’s always some chance that Claude might misinterpret your instructions, you should give Claude very clear guidance around things like this."
Anthropic also warned about the risk of prompt injections, an attack vector in which hackers find a way to disrupt an AI bot's legitimate task with malicious instructions.
That's a threat faced by any AI agent, the company noted, and it's a concern raised by OpenAI too.
FOLLOW US ON SOCIAL MEDIA
Make sure to follow ITPro on Google News to keep tabs on all our latest news, analysis, and reviews.
You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.
Freelance journalist Nicole Kobie first started writing for ITPro in 2007, with bylines in New Scientist, Wired, PC Pro and many more.
Nicole the author of a book about the history of technology, The Long History of the Future.
-
Microsoft Copilot bug saw AI snoop on confidential emails — after it was told not toNews The Copilot bug meant an AI summarizing tool accessed messages in the Sent and Draft folders, dodging policy rules
-
Cyber experts issue warning over new phishing kit that proxies real login pagesNews The Starkiller package offers monthly framework updates and documentation, meaning no technical ability is needed
-
Anthropic promises ‘Opus-level’ reasoning with new Claude Sonnet 4.6 model – and all at a far lower costNews The latest addition to the Claude family is explicitly intended to power AI agents, with pricing and capabilities designed to attract enterprise attention
-
Anthropic reveals Claude Opus 4.6, an enterprise-focused model with 1 million token context window for extended code capabilitiesNews The AI developer highlighted financial and legal tasks, as well as agent tool use, as particular strengths for the new model
-
Why Anthropic sent software stocks into freefallNews Anthropic's sector-specific plugins for Claude Cowork have investors worried about disruption to software and services companies
-
What Anthropic's constitution changes mean for the future of ClaudeNews The developer debates AI consciousness while trying to make Claude chatbot behave better
-
DeepSeek rocked Silicon Valley in January 2025 – one year on it looks set to shake things up again with a powerful new model releaseAnalysis The Chinese AI company sent Silicon Valley into meltdown last year and it could rock the boat again with an upcoming model
-
Workers are wasting half a day each week fixing AI ‘workslop’News Better staff training and understanding of the technology is needed to cut down on AI workslop
-
Anthropic’s Claude AI chatbot is down as company confirms ‘elevated error rates’ for Opus 4.5 and Sonnet 4.5News Users of Anthropic's Sonnet 4.5 and Opus 4.5 models are being met with "elevated error rates"
-
Anthropic announces Claude Opus 4.5, the new AI coding frontrunnerNews The new frontier model is a leap forward for the firm across agentic tool use and resilience against attacks
