Google releases Gemini Canvas API for generative UI components

Tech · 5 min read

The Gemini Canvas API offers endpoints to generate component code, suggest responsive breakpoints, and translate plain-language layout descriptions into proto-components. Google says the API can accept screenshots, design tokens, or plain text and return packages that include React, Flutter, or Web Components implementations plus documentation snippets.

Integration partners are already building plugins for Figma, Android Studio, and Webflow to let designers iterate on ideas faster. Google highlights schema support for existing design systems, allowing teams to provide token manifests and receive components that map to their brand styles. The API also exposes a 'consistency checker' that flags token mismatches and accessibility violations.

Pricing includes a free tier with limited requests and enterprise plans with workspace-level governance. Google emphasizes the API's server-side governance features—fine-grained model version pinning, usage quotas, and content filters—targeted at teams who need predictable, auditable outputs in regulated industries.