Claude Design Now Works in Any Conversation, Including Claude Code

Designing stops being a separate destination

On September 16, 2026, Anthropic merged Claude Cowork and chat into a single Claude and, in the same launch, made Claude Design available from inside any conversation. You no longer pick a mode or open a separate product: you ask for a one-pager, a prototype or a mocked-up onboarding flow, and Claude builds it on a canvas beside the conversation. The Artifacts tab also gains Design templates as a starting point. The standalone experience at claude.ai/design keeps working and existing projects stay where they are.

Claude Design is in beta on the Pro, Max, Team and Enterprise plans and is not available on Free. It is on by default on Pro, Max and Team. On Enterprise it stays off until an owner enables it in organization settings, and the standalone claude.ai/design experience keeps its own separate setting.

What changes inside Claude Code

For Claude Code specifically, design becomes a first-class surface reachable without leaving the terminal:

• /design creates, edits and syncs designs from a Claude Code session, on desktop or in the terminal.
• /design-sync pulls a design system in from your local codebase, so what Claude builds starts from your real components rather than a description of them. Anthropic recommends this path over linking very large repositories directly, which can make the canvas lag.
• A finished design hands off in the other direction too: Send to local coding agent or Send to Claude Code Web continues from the existing work instead of restarting from a screenshot.

That round trip already existed in skeleton form after the June 17 overhaul. What is new is that the design half no longer lives in its own product with its own limits.

Design systems become artifacts

The integrated Claude Design ships a revamped Design Systems feature managed in Settings. Existing team design systems migrate through a banner in the Design tab, and each migrated system becomes an artifact Claude can use in any conversation, including in Claude Code. Anthropic is explicit that migrations are imperfect: each one carries a banner offering to have Claude clean up the guide, the tokens and the components. A Claude Design Admin permission in custom roles lets an admin approve a standard system and lock down edits.

The usage change developers should notice

Claude Design previously had its own weekly allowance, separate from everything else. That allowance is gone. All Claude Design activity now counts toward your plan's shared limits, drawing from the same pool as the rest of your work with Claude, including Claude Code. If you hit your limits, Claude Design is unavailable until they reset, unless you have usage credits enabled. Complex projects with large codebases or many iterations consume more.

This cuts both ways. There is no second budget to track, but heavy design iteration and heavy agentic coding now compete for the same headroom.

Known rough edges

Anthropic ships this as a beta and names the limitations: inline comments occasionally do not render on the page (paste the feedback into chat instead), very large codebases should be linked from Claude Code rather than uploaded, a "chat upstream error" is worked around by opening a new chat tab in the same project, simultaneous multi-person editing is basic and unreliable, and there is no version history yet. On iOS and Android you can ask for a design and view it in the Artifacts tab, but canvas editing and sharing settings need web or desktop.

Why it matters for Claude Code users

The June overhaul made design-to-code a supported handoff. This launch removes the seam entirely: the design system syncs from the repo, the design is produced in the same conversation as the code, the handoff targets a local agent or Claude Code Web, and it all bills against one allowance. For a developer already running Claude Code, the practical effect is that the visual step stops being a detour through another product.