Anthropic Launches Cowork: Claude Desktop Agent Works Directly With Your Files, No Coding Required

United States - Ekhbary News Agency

Anthropic Launches Cowork: Claude Desktop Agent Works Directly With Your Files, No Coding Required

Artificial intelligence leader Anthropic has officially launched Cowork, a significant new capability for its Claude AI agent. This innovative feature extends the power of the company's highly successful Claude Code tool to users without technical backgrounds, enabling them to automate complex tasks directly on their personal computers. Insiders report that the entire feature was developed in an impressively short timeframe of approximately a week and a half, with a substantial portion of the development itself likely utilizing Claude Code. This launch marks a pivotal moment in the race to bring practical AI agents to the mainstream market, significantly bolstering Anthropic's competitive stance against established players like OpenAI and Google in conversational AI, and directly challenging Microsoft's Copilot in the rapidly expanding arena of AI-powered productivity tools.

In an announcement made via its official Claude account on X, the company stated, "Cowork allows you to complete non-technical tasks much like how developers use Claude Code." Currently available as a research preview, access to Cowork is exclusively for Claude Max subscribers, Anthropic's premium tier for power users priced between $100 and $200 per month. The feature is accessible through the macOS desktop application. For the past year, the dominant narrative in the AI industry has centered on large language models adept at creative writing or code debugging. However, Anthropic's Cowork signifies a strategic shift, betting that the true enterprise value lies in an AI capable of performing tangible, real-world tasks – such as organizing a cluttered folder of receipts and generating a structured expense report, all without requiring human intervention.

The genesis of Cowork stems from Anthropic's recent success with its developer community. In late 2024, the company introduced Claude Code, a terminal-based tool designed to help software engineers automate repetitive programming tasks. While Claude Code proved to be a hit, Anthropic observed an unexpected trend: users were repurposing the coding tool for a wide array of non-coding activities. Boris Cherny, an engineer at Anthropic, noted that the company observed users deploying the developer tool for surprisingly diverse tasks. "Since we launched Claude Code, we saw people using it for all sorts of non-coding work: doing vacation research, building slide decks, cleaning up your email, cancelling subscriptions, recovering wedding photos from a hard drive, monitoring plant growth, controlling your oven," Cherny shared on X. He attributed this broad applicability to the underlying strength of the Claude Agent and the Opus 4.5 model, stating, "These use cases are diverse and surprising — the reason is that the underlying Claude Agent is the best agent, and Opus 4.5 is the best model."

Recognizing this emergent pattern of "shadow usage," Anthropic strategically simplified their developer tool, stripping away command-line complexities to create a more accessible, consumer-friendly interface. In their official blog post announcing Cowork, the company explained that developers "quickly began using it for almost everything else," which "prompted us to build Cowork: a simpler way for anyone — not just developers — to work with Claude in the very same way."

The underlying architecture of Cowork is based on an "agentic loop." Unlike a conventional chat interface where users paste text for analysis, Cowork requires a deeper level of trust and access. Users designate a specific folder on their local machine for Claude to access. Within this designated sandbox, the AI agent can read existing files, modify them, or generate entirely new ones. Anthropic provides several practical examples, such as reorganizing a messy downloads folder by intelligently sorting and renaming files, compiling a spreadsheet of expenses from a collection of receipt images, or drafting a report by synthesizing information from scattered notes across multiple documents. "In Cowork, you give Claude access to a folder on your computer. Claude can then read, edit, or create files in that folder," the company elaborated on X. "Try it to create a spreadsheet from a pile of screenshots, or produce a first draft from scattered notes."

When a user assigns a task, the AI doesn't just provide a text-based response. Instead, it formulates a plan, executes the necessary steps, often in parallel, self-corrects its work, and seeks user clarification if it encounters an obstacle. Users can queue multiple tasks, allowing Claude to process them concurrently. Anthropic describes this workflow as feeling "much less like a back-and-forth and much more like leaving messages for a coworker." The system is built upon Anthropic's Claude Agent SDK, sharing the core architecture with Claude Code. This means Cowork can handle many of the same sophisticated tasks as Claude Code, but presented in a more intuitive format suitable for non-coding applications.

Perhaps the most striking aspect of Cowork's launch is the reported speed of its development, highlighting a recursive feedback loop where AI tools are instrumental in building even more advanced AI tools. During a livestream hosted by Dan Shipper, Felix Rieseberg, an Anthropic employee, confirmed the team's rapid development cycle, stating that Cowork was built in approximately a week and a half. AI industry observer Alex Volkov expressed astonishment at this timeline, commenting, "Holy shit Anthropic built 'Cowork' in the last... week and a half?!" This revelation immediately sparked speculation regarding the extent to which Claude Code itself was involved in building Cowork. Simon Smith, EVP of Generative AI at Klick Health, offered a blunt assessment on X: "Claude Code wrote all of Claude Cowork. Can we all agree that we're in at least somewhat of a recursive improvement loop here?" The implication is significant: Anthropic's AI coding agent may have played a substantial role in creating its own non-technical counterpart. If this is indeed the case, it represents one of the most prominent examples to date of AI systems accelerating their own development and expansion, a strategy that could create a considerable advantage for AI labs adept at leveraging internal AI development capabilities.

Cowork's functionality is further enhanced through integration with Anthropic's existing ecosystem of connectors. These tools enable Claude to interface with external information sources and services such as Asana, Notion, and PayPal. Users who have already configured these connections within the standard Claude interface can seamlessly utilize them within Cowork sessions. Moreover, Cowork can be paired with Claude in Chrome, Anthropic's browser extension, to perform tasks that require web interaction. This combination empowers the agent to navigate websites, click buttons, complete forms, and extract data from the internet, all managed through the desktop application. "Cowork includes a number of novel UX and safety features that we think make the product really special," Cherny explained, highlighting "a built-in VM [virtual machine] for isolation, out of the box support for browser automation, support for all your claude.ai data connectors, asking you for clarification when it's unsure." Anthropic has also introduced an initial set of "skills" specifically designed for Cowork, enhancing Claude's ability to create various file types like documents and presentations. These skills build upon the "Skills for Claude" framework announced in October, providing specialized instruction sets that Claude can load for particular task categories.

In a notable display of transparency, Anthropic dedicated significant attention in its announcement to warning users about the potential risks associated with Cowork. The company explicitly acknowledges that Claude "can take potentially destructive actions (such as deleting local files) if it's instructed to." Given that Claude might occasionally misinterpret instructions, Anthropic.