What "One Claude" is and why it merges chat with Cowork
Until now, Claude Cowork (the mode built for longer, more autonomous tasks) and the regular Claude chat lived as separate experiences. With "One Claude," Anthropic puts them in a single app: users no longer have to decide up front whether to open "the chat" or "the work mode" — both capabilities now live in the same place. It's more a product change than a model change — this isn't a new AI, it's a simpler way to reach capabilities Anthropic already had spread across different surfaces.
Claude Docs, Slides and Design: what each one does
Claude Docs lets you build a document collaboratively inside the same conversation: you can edit it directly, add sections, leave comments addressed to Claude, and the document lives at a single shareable link, no exporting required. Claude Slides generates a presentation from any conversation, can be adjusted and presented straight from Claude, and also downloads as PowerPoint or PDF when you need the file outside the platform. Claude Design is the visual piece: mockups, layouts and infographics, without leaving the chat to open another tool.
The pattern across all three is the same: the output stops being a file you have to "pull out" of the AI to keep working on somewhere else, and becomes an object that lives inside the conversation and can keep being edited right there.
Autonomy while you're offline, with the final call still human
Anthropic also described autonomy improvements: Claude can keep working without continuous supervision while the user is offline, with two available modes — confirmation before every action (the default) or continuous operation, only flagging things when it needs attention. The company's own phrase sums up the stance: "the user keeps the final decision." It's the same human-control-over-the-outcome principle showing up in Anthropic's other recent agent-related announcements — autonomy grows in the process, not in approving the result.
Availability, and what it means if you already use Claude for work
The rollout is in beta for Pro and Max plans, across web, desktop and mobile, gradually over several weeks; Enterprise customers will get access with 30 days' advance notice. The announcement adds to a string of 2026 releases from Anthropic — Claude for Teachers in August, for Financial Advisors on September 14, for Small Business on September 15 — in a year where the company also has its Nasdaq IPO planned for October.
For anyone using Claude as a daily work tool — drafting a proposal, putting together a client presentation, getting a mockup ready to discuss — the real advantage isn't a new AI feature. It's removing the step of exporting to Word or PowerPoint just to be able to share it or keep editing it. For a consultant or a small agency, that means less friction between "what Claude generated" and "what the client can actually see and comment on directly."
Talk about your project →