Skip to content

GlamgarOnDiscord/uxui-AI-Prompt

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 

Repository files navigation

uxui-AI-Prompt Banner

uxui-AI-Prompt

The AI skill that builds premium SaaS interfaces — so you don't have to.

Dark-mode first · Animation-driven · Gemini-powered visuals


Stars License

Claude Code Gemini Tailwind Next.js Agent Skills


Get Started · Features · Install · Contributing




⚡ What is this?

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.


🎯 What it does

🎨 Design System

  • Zinc/Slate dark-mode palette — never pure black
  • Geist, Satoshi, Cabinet Grotesk typography
  • Shimmer Buttons, Bento Grids, Liquid Glass Cards
  • WCAG AAA contrast (7:1 normal, 4.5:1 large)
  • Strict anti-patterns enforced automatically

🎛️ 3 Design Dials

DESIGN_VARIANCE  = 8   (symmetry → chaos)
MOTION_INTENSITY = 6   (static → cinematic)
VISUAL_DENSITY   = 4   (gallery → cockpit)

Change a dial in chat → the entire output adapts.

🎬 6 Motion Patterns

Ghost Cursor SVG cursor clicks real UI
Type-Delete-Retype Code rewrites itself
Algorithm Viz Data structure animates
Live Editor Code → preview morphs live
Auto Dashboard KPIs cycle with spring physics
Morphing Metric Stats scramble & cross-fade

🖼️ Gemini Image Generation

Runs automatically at the end of every build:

  1. Audits all image zones in the page
  2. Crafts cinematic prompts per zone
  3. Generates via REST API or Python SDK
  4. Integrates with overlays & alt attrs
  5. Falls back gracefully if no API key

Models: gemini-3.1-flash-image-preview (default) · gemini-3-pro-image-preview (complex scenes)


📁 Architecture

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

🚀 Quick Start

Claude Code — one command:

git clone https://github.com/GlamgarOnDiscord/uxui-AI-Prompt.git
cp -r uxui-AI-Prompt/uxui-designer ~/.claude/skills/uxui-designer

Then 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"     # PowerShell

Zero dependencies (REST API) — works with just curl. No Python needed. Or with Python SDKpip install google-genai Pillow for batch generation.


🔌 Install Anywhere

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

📦 Tech Stack

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

🤝 Contributing

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 → PR

📜 License

Open License — use, modify, and share freely with credit to GlamgarOnDiscord/uxui-AI-Prompt.


Build interfaces that feel alive — with the precision of a senior engineer and the eye of a creative director.

About

My prompt for enhance ur UX UI using AI.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors