Anthropic launches Claude UI, a design critique model tuned for accessibility and tone

AI · 4 min read

Claude UI accepts design artifacts—screenshots, component libraries and copy—and outputs prioritized critique lists covering clarity, microcopy tone, accessibility issues and suggested rewrites. Anthropic emphasized the model’s intent understanding, enabling designers to ask for critiques within specific constraints like ‘brand voice: friendly’ or ‘target audience: seniors’.

The model provides confidence scores and links to guideline references (WCAG, platform HIGs) to make recommendations actionable. Anthropic also added a Slack and Figma plugin so teams can request on-demand design critiques directly in their collaboration channels.

Anthropic is positioning Claude UI as a co-pilot for design reviews rather than an authoritative arbiter; the company recommends human oversight for policy-sensitive areas and offers enterprise controls for how critique data is stored and used.