AI4Figma

About this plugin

This plugin generates high-fidelity UI from text, reference images, and selected frames, then draws fully editable layers with Auto Layout support, effects, and structured hierarchy. Choose mobile or desktop output, generate single screens or full multi-screen flows, and optionally apply design-system styles, reusable components, responsive constraints, and prototype links.Built for speed and control, it includes prompt presets, model/temperature controls, strict JSON mode, accessibility-focused generation, smart edit modes, project-level memory, and an exportable handoff spec with screen stats, tokens, and node summaries.Perfect for ideation, rapid UI exploration, redesign passes, and production-ready starting points directly on your Figma canvas.Available Features:AI UI generation from text promptsReference image–guided generationContext-aware generation from selected existing framesMulti-frame context selection with visual frame cardsOptional replace-in-place for selected framesMobile/Desktop layout selectorSingle-screen and multi-screen flow generationFlow type selection and screen count controlAI-generated design naming for created framesPrompt preset library (Landing, Auth, Dashboard, Profile, Pricing)Model selector (Gemini variants)Temperature controlStrict JSON modeSmart Edit modes (modernize, spacing, accessibility, dark mode)Accessibility mode (WCAG-oriented prompting)Content realism modes (placeholder/product-like/realistic copy)App color scheme editor (primary/background/surface/text + alpha)Live color preview and saved color scheme memoryProject-level settings memory (persisted per file)Design System mode (creates/updates local AI paint styles)Automatic style token mapping to generated nodesReusable Components mode (component + variant set creation)Responsive pass (constraint/resizing hints)Optional flow prototype linking between generated screensLayout QA report (overlap, tap-target, contrast heuristics)Handoff spec export (modal + copy + download JSON)Handoff analytics (per-screen stats, node summary, token extraction, effect usage)