Anthropic launches Claude Design to generate polished visuals fast

Anthropic has just rolled out Claude Design, bringing conversation-first creation for prototypes, slides, one-pagers, and more. Powered by Claude Opus 4.7, it’s a research preview for Pro, Max, Team, and Enterprise with exports and Claude Code handoff.

•
ClaudeAnthropic
claude cover

TL;DR

  • Claude Design (Anthropic Labs): Research preview for polished visuals—designs, prototypes, slides, one-pagers; powered by Claude Opus 4.7
  • Conversation-first iteration: Draft then refine via conversation, inline comments, direct edits, and Claude-generated custom sliders
  • Brand onboarding: Builds a design system from codebase and design files; applies colors, typography, components; supports multiple systems
  • Inputs and capture: Start from prompts, uploads (DOCX/PPTX/XLSX), codebase; web capture tool for existing site elements
  • Collaboration and exports: Org-scoped sharing; private/view/edit options; exports to Canva, PDF, PPTX, standalone HTML
  • Claude Code handoff + controls: Generates handoff bundle for Claude Code; included in Pro/Max/Team/Enterprise; Enterprise off-by-default; extra usage optional; rollout Apr 17, 2026

Claude Design is the latest product to land in Anthropic Labs, framed as a space where Claude can move beyond text and code into polished visual work—including designs, prototypes, slides, and one-pagers. Anthropic says the tool is powered by Claude Opus 4.7 and is launching as a research preview for Claude Pro, Max, Team, and Enterprise subscribers, with a gradual rollout “throughout the day” (dated Apr 17, 2026).

A “conversation-first” path from idea to artifact

The core pitch is straightforward: describe what’s needed, get a first draft, then iterate quickly. Claude Design supports refinement via conversation, inline comments, direct edits, and even custom sliders generated by Claude to adjust things like spacing, color, and layout. Anthropic positions this as useful both for experienced designers who want to explore more directions than time usually allows, and for non-design roles that still need to communicate visual ideas.

That “downstream” handoff story is also explicit: wireframes and mockups created in Claude Design can be handed off to Claude Code for implementation.

How Claude Design handles brand consistency and inputs

Anthropic describes Claude Design as following a “natural creative flow,” with a few concrete mechanics that matter for teams:

  • Brand onboarding: Claude can build a design system by reading a team’s codebase and design files, then apply colors, typography, and components automatically across projects. Teams can refine this over time and maintain more than one system.
  • Import options: Projects can begin from a prompt, or from uploads like DOCX, PPTX, and XLSX, or by pointing Claude at a codebase. There’s also a web capture tool for grabbing elements from an existing website so prototypes resemble the real product.
  • Fine-grained iteration: Inline element-level feedback plus “adjustment knobs,” with Claude applying changes across the design.

Collaboration, exports, and the Claude Code handoff bundle

On the collaboration front, Anthropic says sharing is organization-scoped, with options to keep work private, share via an org link for viewing, or grant edit access so multiple colleagues can edit and chat with Claude in a group thread.

For output, Claude Design can export to Canva, PDF, PPTX, and standalone HTML files, in addition to internal URL sharing and saving as a folder. When it’s time to build, Claude Design can generate a handoff bundle intended for Claude Code, delivered with “a single instruction.”

Availability and admin controls

Claude Design is included for Pro, Max, Team, and Enterprise plans and uses existing subscription limits, with an option to go beyond those limits via extra usage (if enabled). For Enterprise, it ships off by default, and admins can enable it in Organization settings.

Original source: https://www.anthropic.com/news/claude-design-anthropic-labs

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community

Related Articles

claude cover

Opus 4.7 makes Claude Code more autonomous with auto mode

Anthropic’s Opus 4.7 is pushing Claude Code toward longer, more agentic workflows. Boris Cherny details auto mode for fewer permission interruptions, plus recaps, focus mode, effort tuning, and a /go verification loop that can end in a PR.

1 shared tag
claude cover

Anthropic launches Claude Opus 4.7 with stronger coding and vision

Anthropic has just rolled out Claude Opus 4.7, bringing a meaningful upgrade for advanced software engineering and longer agentic runs. It also adds higher-res image support, a new xhigh effort level, API task budgets, and Claude Code’s /ultrareview.

1 shared tag
claude cover

Claude Code’s 1M-token memory can backfire without session discipline

Power users say Claude Code’s million-token context enables longer autonomous workflows—but also invites “context rot” as logs and dead ends pile up. The fix is active session management: rewind, compact with guidance, start fresh, or offload noisy work to subagents.

1 shared tag