The AI skill that builds premium SaaS interfaces — so you don't have to.
Dark-mode first · Animation-driven · Gemini-powered visuals
A plug-and-play AI skill that turns any coding agent into a senior frontend engineer with the eye of a creative director.
Drop it into Claude Code, Cursor, VS Code Copilot, or any compatible tool — and start building interfaces inspired by Vercel, Linear, Stripe, and Raycast in seconds.
Skill mode (
uxui-designer/) — modular, lazy-loaded, optimized for Claude Code & compatible agents. Prompt mode (prompt.md) — single-file system prompt for ChatGPT, Windsurf, Aider, or anything else.
|
Change a dial in chat → the entire output adapts. |
||||||||||||||
|
Runs automatically at the end of every build:
Models: |
uxui-AI-Prompt/
│
├── prompt.md ← Standalone prompt (any AI tool)
│
└── uxui-designer/ ← Skill (Claude Code & compatible agents)
├── SKILL.md ← Entry point — loads references on demand
└── references/
├── design-system.md ← Colors · Typography · Layout · Components
├── motion-patterns.md ← Framer Motion · GSAP · 6 autonomous demos
├── copywriting.md ← Copy rules · CTAs · Anti-patterns
├── page-structure.md ← 8 mandatory sections with specs
└── image-generator.md ← Gemini API pipeline (REST + SDK)
User prompt → SKILL.md → lazy-load references → build page → image-generator → done
Claude Code — one command:
git clone https://github.com/GlamgarOnDiscord/uxui-AI-Prompt.git
cp -r uxui-AI-Prompt/uxui-designer ~/.claude/skills/uxui-designerThen just ask Claude to build any UI:
Build a landing page for a developer analytics SaaS.
Next.js + Tailwind. Dark mode. Emphasize the real-time dashboard.
The skill handles everything: onboarding questions → design dials → layout → animations → Gemini visuals.
Optional — Enable Gemini image generation
Get a free API key at aistudio.google.com/apikey, then:
export GEMINI_API_KEY="your-key-here" # Linux / macOS
$env:GEMINI_API_KEY = "your-key-here" # PowerShellZero dependencies (REST API) — works with just curl. No Python needed.
Or with Python SDK — pip install google-genai Pillow for batch generation.
Claude Code
| Scope | Command |
|---|---|
| Personal (all projects) | cp -r uxui-designer ~/.claude/skills/uxui-designer |
| Project (team-shared) | mkdir -p .claude/skills && cp -r uxui-designer .claude/skills/uxui-designer |
| Session (temporary) | claude --add-dir /path/to/uxui-designer |
Cursor · VS Code Copilot · Amp · Junie · Goose
Copy the uxui-designer/ folder into your tool's skill directory:
| Tool | Path |
|---|---|
| Cursor | .cursor/skills/uxui-designer/ |
| VS Code Copilot | .github/skills/uxui-designer/ |
| Amp | .amp/skills/uxui-designer/ |
| Junie (JetBrains) | .junie/skills/uxui-designer/ |
| Goose | .goose/skills/uxui-designer/ |
Follows the open Agent Skills standard — any compatible tool auto-discovers
SKILL.md.
Windsurf · Aider · ChatGPT · Others
These tools don't support skills — use the standalone prompt instead:
| Tool | How |
|---|---|
| Windsurf | Paste prompt.md contents into .windsurfrules |
| Aider | aider --read prompt.md |
| ChatGPT / Other | Paste prompt.md as system prompt |
| Framework | React · Next.js App Router (preferred) · Static HTML |
| Styling | Tailwind CSS |
| Animation | Framer Motion · GSAP / ScrollTrigger · Tailwind Animate |
| Icons | Phosphor Icons (preferred) · Lucide · HugeIcons |
| Components | Magic UI · Aceternity UI · ShadCN · Reactbits |
| Images | Gemini gemini-3.1-flash-image-preview / gemini-3-pro-image-preview |
PRs, ideas, and issues welcome. You can improve the design system, add motion patterns, or refine the Gemini prompting logic.
fork → git checkout -b my-feature → PROpen License — use, modify, and share freely with credit to GlamgarOnDiscord/uxui-AI-Prompt.