Claude chat absorbs Cowork and adds Docs and Slides

Anthropic is merging its Claude chat interface with Claude Cowork, a workspace that lets Claude execute multistep jobs using approved files, connectors, and tools. Once the rollout reaches an account, any conversation can become a cloud task that continues after the user closes the tab or shuts down a laptop. Claude Docs and Claude Slides are also launching in beta, while Claude Design is moving into conversations.

The merger removes the need to choose between chat and Cowork before starting a task. Context, connected services, reusable skills, and generated files can remain attached to one conversation as the work expands. The announcement covers Claude’s user-facing apps; API changes and Claude Code are outside its scope.

Pro and Max go first

Anthropic’s announcement says the combined experience will reach Pro and Max subscribers over the next few weeks. Other plans will follow in stages.

Claude Docs, Slides, and Design are beta features for paid plans. Enterprise administrators can decide when to enable them, and the standalone version of Claude Design will remain available.

Claude chooses the execution path

Anthropic launched Cowork in January 2026 as a research preview that adapted Claude Code’s agent workflow for knowledge workers. The product developed into a workspace with projects, connectors, plug-ins, scheduled runs, and cloud execution, but users had to classify each request as a chat or Cowork task before beginning.

In the combined interface, Claude can answer a straightforward prompt immediately or launch a longer job from the same thread. Connectors provide access to authorized external services, while skills package reusable instructions and workflows. Both remain available as a task grows in scope.

Claude will continue to request permission before taking actions by default. Users can enable a setting that allows work to continue until the agent encounters an action requiring review. Existing Cowork conversations, projects, artifacts, connectors, and skills remain in place.

Editable files stay in the thread

Claude Docs and Claude Slides add editable documents and presentations to the conversation. Claude Design provides a matching surface for visual assets. The three apps support a workflow that moves from drafting to editing, review, presentation, and export without leaving Claude.

• Draft a one-page document, then convert it into a presentation from the same conversation.
• Create a matching visual in Design using the existing context.
• Edit generated content directly or leave comments for Claude to address.
• Select, move, and revise individual elements through direct manipulation or written instructions.
• Present from Claude, export to PowerPoint or PDF, or share the work through one link.

These outputs function as persistent artifacts rather than static chat attachments. Teammates can review them, add comments, and continue working from the conversation that produced them.

Cloud execution closes the gap

Standard Claude chat historically responded to messages without direct access to local files. Cowork could read, edit, and create files inside folders selected by the user, giving the agent enough access to complete tasks. That permission model originally justified a separate workspace.

Cowork later moved execution into Anthropic’s cloud environment, where the task runs independently of the user’s device. Sessions can start from claude.ai, the mobile app, or Claude Desktop and continue after the initiating device disconnects. Extending that runtime to ordinary conversations removes the operational reason for a separate entry point.

The unified interface also makes Anthropic’s expanding connector catalog available from a regular conversation. Supported services include Google Workspace, Microsoft 365, Slack, HubSpot, QuickBooks, PayPal, Canva, and DocuSign. The design more closely resembles ChatGPT’s primary app, which brings canvases, tasks, and connectors into shared threads.

Recurring work fits in one conversation

A weekly report can now become a scheduled cloud task that gathers information through approved connectors, drafts a document every Monday, accepts teammate comments, and generates a presentation from the same context. The resulting files can remain attached to one shareable conversation.

Existing Cowork projects carry into the merged interface, preserving their connectors, skills, files, and schedules. During the beta, Anthropic is concentrating document creation, visual design, review, exports, and agent actions in the same workspace, making the conversation a persistent record of both the instructions and the finished work.