Rakit UI AI
An intelligent tool for AI assistants to present multiple UI component designs for user selection.
Rakit UI AI - MCP Design Selection Tool
An intelligent UI component design selection tool built on the Model Context Protocol (MCP). Generate and compare UI designs using AI, or provide your own designs for selection.
Features
- π€ AI-Powered Generation - Generate designs from natural language using MiniMax-M2.1
- π¨ Visual Comparison - Side-by-side design comparison in a modern browser interface
- π± Responsive Preview - Mobile, desktop, and responsive view modes
- π HTML Reference Analysis - Automatically checks design completeness
- π― Purpose-Driven - Specify what to include/exclude in designs
- β‘ Real-time Selection - WebSocket support for instant feedback
Quick Start
# Install globally
npm install -g rakitui-ai
# Set MiniMax API key (for AI generation)
export MINIMAX_API_KEY="your-api-key-here"
# Run the MCP server
rakitui-ai
Setup
Cursor IDE
- Open Settings (
Cmd+,orCtrl+,) - Navigate to MCP settings
- Add new server:
- Name:
rakitui-ai - Command:
rakitui-ai - Environment Variables:
MINIMAX_API_KEY(optional)
- Name:
Or edit settings.json:
{
"mcpServers": {
"rakitui-ai": {
"command": "rakitui-ai",
"env": {
"MINIMAX_API_KEY": "your-api-key-here"
}
}
}
}
- Restart Cursor
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"rakitui-ai": {
"command": "rakitui-ai"
}
}
}
Usage
AI-Powered Generation
Simply describe what you need:
const result = await tool.execute({
prompt: "Create 3 modern button designs for a SaaS dashboard",
style_preference: "clean and professional",
framework: "tailwind",
component_type: "button",
auto_search_reference: true, // Get design references
purpose_include: ["hover state", "focus state"], // Must include
purpose_exclude: ["animations"], // Must avoid
view_mode: "responsive" // mobile | desktop | responsive
});
Manual Design Input
Provide your own designs:
const result = await tool.execute({
design_name_1: "Modern Button",
design_html_1: `<button class="bg-blue-500 text-white px-4 py-2 rounded">Click me</button>`,
design_name_2: "Classic Button",
design_html_2: `<button class="btn btn-primary">Click me</button>`,
design_name_3: "Minimal Button",
design_html_3: `<button class="px-4 py-2 border rounded">Click me</button>`
});
Response
{
"success": true,
"selectedDesign": "Modern Button",
"selectedDesignHtml": "<button>...</button>",
"design_options": [...],
"selection_context": {
"purpose": {...},
"references": [...],
"view_mode": "responsive"
}
}
Parameters
AI Generation Parameters
| Parameter | Type | Description |
|---|---|---|
prompt | string | Natural language description of the UI component |
style_preference | string | Style guidance (e.g., "modern", "minimalist") |
framework | string | CSS framework (tailwind, bootstrap, bulma, etc.) |
component_type | string | Component type (button, card, form, navigation, etc.) |
auto_search_reference | boolean | Enable automatic reference search |
purpose_include | string[] | Elements that must be included |
purpose_exclude | string[] | Elements to avoid |
purpose_notes | string | Additional requirements |
view_mode | string | Preview mode: "mobile", "desktop", or "responsive" |
Manual Input Parameters
| Parameter | Type | Description |
|---|---|---|
design_name_1/2/3 | string | Name of each design |
design_html_1/2/3 | string | HTML content of each design |
Supported Frameworks
- Tailwind CSS
- Bootstrap
- Bulma
- Foundation
- Semantic UI
- Plain CSS
Supported Components
- Buttons (primary, secondary, outlined, with icons)
- Cards (content, featured, compact)
- Forms (inputs, validation, layouts)
- Navigation (bars, menus, responsive)
- Modals (dialogs, confirmations)
- Tables (data, striped, responsive)
Keyboard Shortcuts
1,2,3- Quick select designEscape- Close zoom view- Click design - Zoom in for detailed view
Troubleshooting
Command not found:
- Ensure npm global bin is in your PATH
- Or use full path:
/usr/local/bin/rakitui-ai(Mac)
Tool not appearing:
- Restart your IDE completely
- Check MCP server status indicator
MiniMax API not working:
- Verify
MINIMAX_API_KEYis set correctly - Check your MiniMax account has credits
Development
# Clone and install
git clone https://github.com/madebyaris/rakitui-ai.git
cd rakitui-ai
npm install
# Build
npm run build
# Start
npm start
License
MIT License - see LICENSE file for details.
Links
- Repository: github.com/madebyaris/rakitui-ai
- Issues: GitHub Issues
- MiniMax API: platform.minimax.io
Server Terkait
Alpha Vantage MCP Server
sponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Algorand
A comprehensive MCP server for tooling interactions(40+) and resource accessibility(60+) plus many useful prompts to interact with Algorand Blockchain.
MetaTrader 4
Integrate with the MetaTrader 4 trading platform to access trading functions and data via an HTTP bridge and Expert Advisor.
Remote MCP Server (Authless)
An example of a remote MCP server deployable on Cloudflare Workers, without authentication.
WordPress Community DEV Docs
Access WordPress development rules and best practices from the WordPress LLM Rules repository. It dynamically creates tools for each rule and caches content using Cloudflare Durable Objects.
TradingCyborg MCP Server
A professional trading server with over 26 tools for Bybit API integration.
CodeRabbit
Integrate with CodeRabbit AI for automated code reviews, pull request analysis, and report generation.
AI Design Blueprint Doctrine
The industry-standard doctrine for safe, observable, and steerable AI agent UX β browse 10 principles, curated examples, and application guides via MCP.
PawSift πΎ for Android Logcat
PawSift bridges Android Logcat to LLMs in a token-efficient way
Recent Go MCP Server
Provides Go language updates and best practices in a structured Markdown format for LLM coding agents.
mcp-of-mcps
MCP of MCPs is a meta-server that merges all your MCP servers into a single smart endpoint.β¨It gives AI agents instant tool discovery, selective schema loading, and massively cheaper execution, so you stop wasting tokens and time. With persistent tool metadata, semantic search, and direct code execution between tools, it turns chaotic multi-server setups into a fast, efficient, hallucination-free workflow.β¨It also automatically analyzes the tools output schemas if not exist and preserves them across sessions for consistent behavior.