Anthropic launches Claude Design, a safety-tuned model for brand-sensitive UI generation

AI · 5 min read

Claude Design is a tuned descendant of Anthropic's Claude family and incorporates policy-aware decoding to reduce hallucinated brand claims, misleading UIs, and offensive content. The model supports multimodal prompts including screenshots, brand tokens, and legal constraints.

Anthropic emphasized auditability: each generation returns a rationale trace that explains why certain assets or copy were suggested, along with confidence scores and policy checks. The company says this helps compliance teams review AI-produced UI copy and flows faster.

Claude Design is available via Anthropic's enterprise API with per-tenant custom tuning and data handling options. Pricing and SLA tiers are targeted at financial services, healthcare, and public sector customers that require strict data governance.