Anthropic wants to expand the capabilities of Claude’s AI agent and capitalize on the growing buzz around Claude Code — and it’s doing so with a brand up-to-date feature released on Monday called “Claude Cowork.”
“Cowork can do many of the same things Claude Code can do, but in a more accessible form for non-coding tasks,” Anthropic wrote in her article blog post. The company is releasing it as a “preview” so the team can learn more about how people utilize it and continue working on it. For now, Cowork is only available through the Claude app on macOS and only to subscribers of Anthropic’s advanced user tier, Claude Max, which costs between $100 and $200 per month depending on usage.
Here’s how Claude Cowork works: The user gives Claude access to a folder on their computer, allowing the chatbot to read, edit, and create files. (Examples Anthropic gave included the ability to “reorganize downloaded files by sorting and renaming each file, create a new spreadsheet listing expenses from a stack of screenshots, or create a first draft of a report from scattered notes.”) Claude will provide regular updates on what he’s working on, and users can also utilize existing connectors to connect it to external information (such as Asana, Notion, PayPal and other supported partners) or connect it to Claude in Chrome to perform browser-related tasks.
“You don’t have to manually provide context or convert Claude’s results to the appropriate format,” Anthropic wrote. “You also don’t have to wait for Claude to finish before offering further ideas or comments: you can queue up tasks and let Claude work on them in parallel. It’s much less like going back and forth and more like leaving a message with a colleague.”
The up-to-date feature is part of Anthropic’s (and its competitors’) drive to provide the most useful AI agents for both consumers and enterprises. AI agents have come a long way from their humble beginnings as mostly theoretically useful tools, but a lot more work is needed before you see your non-tech friends using them to complete everyday tasks.
Anthropic’s “Skills for Claude” program, announced in October, was a partial precursor to Cowork. Starting in October, Claude can improve on personalized tasks and tasks thanks to “folders containing instructions, scripts and resources that Claude can load as needed to smarter perform specific work tasks – from working with Excel [to] in line with your organization’s branding guidelines,” according to the publication at the time. People could also build their own Skills for Claude in relation to their specific positions and the tasks they needed to complete.
As part of the announcement, Anthropic warned of the potential risks of using Cowork and other AI agent tools – namely that if instructions are not clear, Claude could delete local files and take other “potentially destructive actions” – and that there are a number of potential security risks in the case of instantaneous injection attacks. Instant injection attacks often involve bad actors hiding malicious text on a website referenced by the model, which instructs the model to bypass security and do something malicious, such as handing out personal information. “Agent security – or the task of securing Claude’s real-world activities – continues to be an active area of development in the industry,” Anthropic wrote.
Claude Max subscribers will try the up-to-date feature by clicking “Cowork” in the sidebar of the macOS app. Other users can join waiting list.