Skip to main content
Glama
mcp-claude-design-arabic

mcp-claude-design-arabic

MCP Claude Design Arabic 🎨🇸🇦

npm version License: MIT Node.js Version MCP Protocol Author

An elite, production-ready Model Context Protocol (MCP) server engineered specifically for modern Arabic and bilingual (RTL/LTR) UI/UX design, interactive agent cognition, anti-AI-slop detection, and Level-5 Claude visual design engineering.


🌍 Language Navigation / التنقل حسب اللغة


🇬🇧 English Documentation

Related MCP server: Font MCP

🚀 Overview

MCP Claude Design Arabic (v2.0.0) by MarwanDevSpace bridges the gap between AI code generation and authentic, ultra-polished Arabic/Bilingual UI/UX design. Inspired by production design workspaces like open-design, it combines interactive design agent personas (ArabicArtDirector, RTLArchitect, MicroInteractionEngineer, BilingualDesignStrategist, CognitiveUXAuditor), dynamic RTL/LTR logical property conversion, 3-dial aesthetic configuration (VARIANCE, MOTION, DENSITY), anti-AI-slop audit checks, optical typography alignment, and Claude Level-5 visual design benchmarks into a single high-performance MCP server.


🌟 Key Capabilities

  • Intuitive Tool Calling (pipeline_design): Sequentially orchestrates Brief Inference (Design Read), Art Direction, RTL property & BiDi form mapping, typography metric analysis, micro-animation curves, ARIA accessibility, Anti-Slop auditing, and pre-flight checklist evaluation in a single tool call.

  • Design Taste Dials & Brief Inference: Automatically infers 3 aesthetic dials (DESIGN_VARIANCE, MOTION_INTENSITY, VISUAL_DENSITY) from prompt context to customize grid asymmetry, motion physics, and visual density.

  • Anti-AI-Slop Auditor & Pre-Flight Matrix (anti_slop_audit): Detects and flags 7 cardinal sins (default indigo hexes #6366f1, two-stop trust gradients, raw emoji feature icons, positive Arabic tracking, rounded cards with left borders, fake metrics, filler copy) and strictly enforces the Em-Dash Ban (—).

  • Bi-Directional RTL/LTR & Form BiDi Engine (rtl_layout_matrix): Automatically refactors physical CSS/Tailwind/Flutter properties (ml-*, pr-*) to logical properties (ms-*, pe-*), applies 180° rotation (rtl:rotate-180) to directional icons, and enforces dir="auto" for free text and forced dir="ltr" for email/phone/url inputs.

  • Arabic Typography & Metric Compensator (arabic_typography): Enforces CJK/Arabic display line-height rules (1.35x - 1.45x), body line-height (1.65x - 1.85x), baseline shifts, 65ch line length, and the Three-Weight System (400 Read, 550 Emphasize, 600 Announce).

  • Claude Level-5 Benchmark Scoring (evaluate_benchmark): Evaluates UI code against 6 quality axes: Typography, Glassmorphic Depth, RTL Fluidity, Micro-Animations, Accessibility, and Anti-Slop Discipline.


🛠️ Tool Invocation Reference

Primary Tool

Short Alias

Parameters

Description

pipeline_design

run_full_design_pipeline

componentCode, brandName, primaryColor, accentColor, arabicFont, targetFormat

Executes full multi-agent design thinking pipeline and returns refactored UI code

anti_slop_audit

—

codeSnippet

Audits UI code for anti-AI-slop violations, em-dash bans, rounded left-border cards, and fake metrics

design_tokens

generate_claude_design_tokens

brandName, primaryColor, accentColor, themeMode, format

Generates OKLCH color palettes, glassmorphism tokens, semantic layer variables, and motion curves

rtl_layout_matrix

generate_rtl_ltr_layout_matrix

componentSnippet, targetFormat, flipDirectionalIcons

Converts physical layout properties to logical RTL/LTR properties, flips directional icons, and applies form BiDi rules

arabic_typography

analyze_arabic_typography

textSnippet, currentFont, fontSizePx, hasDiacritics, targetPlatform

Audits Arabic font metrics, line-height compensation, baseline shifts, and Latin font pairings

accessibility_audit

audit_bilingual_accessibility

markup, checkContrast, checkBidiSemantics

Audits markup for dir="rtl", lang="ar", <bdi> isolation, Touch Targets (24/44px), and ARIA landmarks

evaluate_benchmark

—

codeOrMarkup, context, strictMode

Scores UI code against Level-5 Claude Design standards (0-100) with detailed subscores and anti-slop violations

invoke_design_agent

invoke_interactive_design_agent

agentRole, prompt, projectContext, codeSnippet

Consults individual AI agent personas for domain-specific design critique and code overlays


💬 MCP Prompts & Shortcuts

Invoke prompt templates directly in your AI editor:

  • /pipeline_design or /arabic_ui_masterpiece: System prompt guiding AI to craft luxury Level-5 Arabic UI components.

  • /creative_direction_brief: Generates a Creative Director Brief (Step 0) defining Design Read, taste dials, and anti-patterns.

  • /anti_slop_audit: Generates an Anti-AI-Slop audit prompt checking code against the 7 cardinal sins & em-dash ban.

  • /pre_flight_checklist: Generates a 20-point mechanical pre-flight audit checklist before publishing.

  • /rtl_layout_conversion: Refactoring instructions for converting legacy LTR code to logical RTL/LTR.


⚙️ Environment Setup Guide

1. Cursor / Antigravity Setup

Add mcp-claude-design-arabic to your mcp.json configuration file:

{
  "mcpServers": {
    "claude-design-arabic": {
      "command": "npx",
      "args": [
        "-y",
        "--silent",
        "mcp-claude-design-arabic@latest"
      ]
    }
  }
}

2. Claude Code & Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "claude-design-arabic": {
      "command": "npx",
      "args": [
        "-y",
        "--silent",
        "mcp-claude-design-arabic@latest"
      ]
    }
  }
}

3. OpenCode / Codex / Windsurf / VS Code

{
  "mcpServers": {
    "claude-design-arabic": {
      "command": "npx",
      "args": [
        "-y",
        "--silent",
        "mcp-claude-design-arabic@latest"
      ]
    }
  }
}


🇸🇦 الوثائق باللغة العربية

🚀 نظرة عامة

يقدم MCP Claude Design Arabic (الإصدار v2.0.0) المطور بواسطة MarwanDevSpace خوارزمية ذكية وشاملة مخصصة لتصميم وتطوير الواجهات وتطبيقات الويب العربية وثنائية الاتجاه (RTL/LTR) بمستوى جودة Claude Level-5 Design Standards.

يستغل السيرفر قوة المعرفة المكتسبة من مكتبة التصميم المحترفة open-design ويجمع بين 5 وكلاء تصميم تفاعليين (ArabicArtDirector, RTLArchitect, MicroInteractionEngineer, BilingualDesignStrategist, CognitiveUXAuditor) مع محرك تحويل الاتجاهات اللوجستية، كاشف التصميم السطحي (Anti-AI-Slop)، ومحرك الذوق البصري الثلاثي.


🌟 الميزات الأساسية

  • أمر النداء المباشر والسلس (pipeline_design): تشغيل تسلسلي ذكي للخدمات في أداة واحدة (التشخيص البصري ➔ التوجيه الفني ➔ التحويل المنطقي والتوجيه التلقائي للحقول ➔ ضبط الخطوط ➔ الحركات الدقيقة ➔ كشف الذكاء الاصطناعي السطحي ➔ تدقيق الإتاحة ➔ مصفوفة ما قبل النشر).

  • معدّلات الذوق البصري وتدقيق AI السطحي (anti_slop_audit): تحديد تلقائي لمقاييس التباين والتفاعلات، وكشف الذنوب السبعة المرفوضة (التدرجات الأرجوانية السطحية #6366f1، الإيموجي في الأيقونات، وحظر شرطة Em-Dash — تماماً).

  • محرك التحويل ثنائي الاتجاه وحقول BiDi (rtl_layout_matrix): استبدال تلقائي للهوامش والمحاذاة الفيزيائية بالخصائص المنطقية (ms-*, pe-*, start-*) وعكس الأيقونات الاتجاهية تلقائياً (rtl:rotate-180) وفرض dir="auto" و dir="ltr" على حقول النموذج.

  • محرك طباعة الخطوط العربية (arabic_typography): ضبط ارتفاع السطر (1.65x - 1.85x للنصوص، 1.35x - 1.45x للعناوين) وتطبيق نظام الأوزان الثلاثي وقاعدة طول السطر 65ch.


🛠️ جدول الأدوات وأوامر المناداة المباشرة

الأداة الأساسية

اسم الاستدعاء المختصر

الوصف

pipeline_design

run_full_design_pipeline

تشغيل خط تصميم الوكلاء التسلسلي الموحد وإرجاع الكود المحسن بالكامل

anti_slop_audit

—

كشف وتدقيق الواجهة ضد مؤشرات الذكاء الاصطناعي السطحي وشرطة Em-Dash

design_tokens

generate_claude_design_tokens

توليد توكنات تصميم متكاملة (ألوان OKLCH، أسطح شفافة، توكنات دلالية)

rtl_layout_matrix

generate_rtl_ltr_layout_matrix

تحويل التنسيق الفيزيائي إلى كلاسات منطقية وعكس الأيقونات وتطبيق قواعد BiDi للحقول

arabic_typography

analyze_arabic_typography

تحليل الخط العربي وارتفاع السطر وتطبيق أوزان القراءة والدعوة للإجراء

accessibility_audit

audit_bilingual_accessibility

تدقيق سهولة الوصول، تباين WCAG 2.2 AA، وأهداف اللمس (24/44px) وعزل <bdi>

evaluate_benchmark

—

تقييم كود الواجهة وفق معايير التصميم الفاخرة وكشف مؤشرات التصميم السطحي

invoke_design_agent

invoke_interactive_design_agent

استشارة أحد وكلاء التصميم التفاعليين للحصول على توجيهات وتراكيب كود مخصصة


⚙️ دليل التثبيت والربط في المحررات والبيئات

التثبيت الموحد عبر mcp.json أو claude_desktop_config.json

أضف السيرفر مباشرة في محررك المفضل (Cursor / Antigravity / Claude Code / Windsurf / VS Code / Codex):

{
  "mcpServers": {
    "claude-design-arabic": {
      "command": "npx",
      "args": [
        "-y",
        "--silent",
        "mcp-claude-design-arabic@latest"
      ]
    }
  }
}

📄 License

MIT © MarwanDevSpace

Available Tools

14 tools
accessibility_auditB

Audits HTML/JSX code markup for Arabic screen reader compatibility (dir="rtl", lang="ar"), BiDi text isolation, touch targets (24px/44px), focus rings, and WCAG 2.2 AA compliance.

Inputs: markup, checkContrast, checkBidiSemantics.

analyze_arabic_typographyA

Audits an Arabic or bilingual text snippet for optical alignment, font pairing, line-height compensation, and ligature integrity.

Inputs: fontSizePx, currentFont, textSnippet, hasDiacritics, +1 more.

anti_slop_auditA

Audits UI code for anti-AI-slop violations, em-dash bans, rounded left-border cards, and fake metrics.

Inputs: codeSnippet (required).

arabic_typographyB

Audits an Arabic or bilingual text snippet for optical alignment, font pairing, line-height compensation, and ligature integrity.

Inputs: fontSizePx, currentFont, textSnippet, hasDiacritics, +1 more.

audit_bilingual_accessibilityA

Audits HTML/JSX code markup for Arabic screen reader compatibility (dir="rtl", lang="ar"), BiDi text isolation, touch targets (24px/44px), focus rings, and WCAG 2.2 AA compliance.

Inputs: markup, checkContrast, checkBidiSemantics.

design_tokensC

Generates ultra-premium Claude Level-5 design tokens (OKLCH color palettes, semantic layers, glassmorphism shadows, motion physics, Arabic typography scales) for Tailwind v4, CSS, JSON, or Flutter.

Inputs: format, brandName, themeMode, accentColor, +1 more.

evaluate_design_benchmarkA

Evaluates UI code, CSS, or Tailwind markup against Level-5 Claude Design standards, returning overall score, sub-scores, strengths, and actionable visual polish recommendations.

Inputs: context, strictMode, codeOrMarkup.

generate_claude_design_tokensB

Generates ultra-premium Claude Level-5 design tokens (OKLCH color palettes, semantic layers, glassmorphism shadows, motion physics, Arabic typography scales) for Tailwind v4, CSS, JSON, or Flutter.

Inputs: format, brandName, themeMode, accentColor, +1 more.

generate_rtl_ltr_layout_matrixB

Converts legacy physical CSS/Tailwind/Flutter component code into modern bidirectional (RTL/LTR) logical components with directional icon flip and Form BiDi rules.

Inputs: targetFormat, componentSnippet, primaryDirection, flipDirectionalIcons.

invoke_design_agentA

Dispatches design challenges to specialized adaptive AI design agents (ArabicArtDirector, RTLArchitect, MicroInteractionEngineer, BilingualDesignStrategist, CognitiveUXAuditor) to get deep design critique, strategic guidance, and code optimizations.

Inputs: prompt, agentRole, codeSnippet, projectContext.

invoke_interactive_design_agentB

Dispatches design challenges to specialized adaptive AI design agents (ArabicArtDirector, RTLArchitect, MicroInteractionEngineer, BilingualDesignStrategist, CognitiveUXAuditor) to get deep design critique, strategic guidance, and code optimizations.

Inputs: prompt, agentRole, codeSnippet, projectContext.

pipeline_designC

Executes the unified multi-agent design pipeline (Brief Inference -> Art Direction -> RTL & BiDi Forms -> Arabic Typography -> Micro-Interactions -> Anti-Slop Audit -> Pre-Flight Checklist -> Benchmark) to generate clean, production-ready Arabic/Bilingual UI code.

Inputs: brandName, arabicFont, accentColor, primaryColor, +2 more.

rtl_layout_matrixB

Converts legacy physical CSS/Tailwind/Flutter component code into modern bidirectional (RTL/LTR) logical components with directional icon flip and Form BiDi rules.

Inputs: targetFormat, componentSnippet, primaryDirection, flipDirectionalIcons.

run_full_design_pipelineB

Executes the unified multi-agent design pipeline (Brief Inference -> Art Direction -> RTL & BiDi Forms -> Arabic Typography -> Micro-Interactions -> Anti-Slop Audit -> Pre-Flight Checklist -> Benchmark) to generate clean, production-ready Arabic/Bilingual UI code.

Inputs: brandName, arabicFont, accentColor, primaryColor, +2 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3/5.0

    Scored across 14 tools

    Disambiguation1/5

    Every unique tool has an exact duplicate with a different name but identical description (e.g., pipeline_design and run_full_design_pipeline). An agent cannot distinguish between them, leading to misselection and confusion. This is the worst possible overlap.

    Naming Consistency2/5

    Naming conventions are inconsistent: some tools use verb-first patterns (analyze_arabic_typography, generate_design_tokens) while others are noun-first (arabic_typography, design_tokens). The same functionality uses different naming styles (pipeline_design vs run_full_design_pipeline), making the API unpredictable.

    Tool Count4/5

    14 tools is within the expected range for a design pipeline server, but the count is inflated by exact duplicates. With only 7 unique tools, the effective surface is well-scoped, though the redundancy makes the actual count slightly higher than ideal.

    Completeness4/5

    The unique tools cover the core domain: pipeline execution, audits (visual, accessibility, typography), layout conversion, benchmark evaluation, design agents, and token generation. Minor gaps exist, such as standalone tools for individual pipeline stages, but the pipeline and agent tools provide workarounds.

    Maintenance

    ActivityNo data
    ResponsivenessSyncing

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      MCP server for VertaaUX.ai — run UX & accessibility audits, generate fixes, and monitor quality from your LLM or IDE. 38 tools covering the full audit-to-fix lifecycle across 7 UX categories.
      29
      MIT
    • F
      license
      Not graded
      quality
      C
      maintenance
      An expert design advisor MCP server that researches live discussions, scans project files, and automatically downloads font files for your project setup.
      10
      -
    • A
      license
      A
      quality
      C
      maintenance
      MCP server for correct Arabic formatting — currency, Hijri dates, number-to-words, RTL fixes and validation across all 22 Arab countries. Zero-dependency.
      17
      141 npm
      3
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      MCP server that bridges design and code by generating UI components, layouts, and themes from text prompts using v0.dev, with features like code refactoring and accessibility auditing.
      1
      MIT