MCP ExplorerExplorer

Mcp Server Remix Icon

@FradSeron a year ago
4 MIT
FreeCommunity
AI Systems
Remix Icon MCP is an advanced icon search and recommendation service that leverages Cloudflare Workers. It offers smart icon discovery using natural language descriptions, multi-language support, and category management, powered by various semantic matching algorithms for optimal search results.

Overview

What is Mcp Server Remix Icon

mcp-server-remix-icon is a powerful icon search and recommendation service built on Cloudflare Workers, utilizing advanced semantic matching algorithms for intelligent icon discovery.

Use cases

Use cases include integrating icon search in design software, enhancing user experience in web applications with icon recommendations, and providing a resource for developers needing quick access to icons based on descriptions.

How to use

To use mcp-server-remix-icon, you can call its API endpoints such as ‘findIcons’ to search for icons based on descriptions, or ‘getIconCategories’ to retrieve available icon categories.

Key features

Key features include smart icon search using natural language, multi-language support for English and Chinese, category management, advanced matching algorithms, inverted index for fast search, and LRU caching for performance optimization.

Where to use

mcp-server-remix-icon can be used in web applications, design tools, and any platform requiring efficient icon search and management functionalities.

Content

Remix Icon MCP

Twitter Follow

English | 简体中文

A powerful icon search and recommendation service built on Cloudflare Workers, providing intelligent icon discovery through advanced semantic matching algorithms.

Features

  • Smart Icon Search: Find icons based on natural language descriptions using multiple similarity algorithms
  • Multi-language Support: Optimized for both English and Chinese text input
  • Category Management: Browse and search icons by categories
  • Advanced Matching: Uses multiple algorithms for better search results:
    • Jaccard Similarity
    • N-gram Matching
    • Category Matching
    • Exact Matching
    • Levenshtein Distance
    • Name Matching
    • Tag-based Matching
  • Inverted Index: Fast preliminary search using an inverted index
  • Caching: LRU caching for improved performance

API Endpoints

Find Icons

findIcons(description: string): ResponseContent[]

Finds icons based on user description, returns top 5 recommendations with similarity scores.

Get Icon Categories

getIconCategories(): ResponseContent[]

Returns a list of all available icon categories.

Find Icons by Category

findIconsByCategory(description: string, category: string): ResponseContent[]

Searches for icons within a specific category based on description, returns top 5 recommendations.

Project Structure

.
├── src/                   # Source code directory
│   ├── index.ts           # Main entry point
│   ├── data/              # Data files including icon catalog
│   ├── domain/            # Domain models and services
│   │   ├── icon/          # Icon domain models
│   │   └── search/        # Search functionality
│   ├── infrastructure/    # Infrastructure components
│   │   ├── logging/       # Logging utilities
│   │   └── result/        # Result handling
│   └── utils/             # Utility functions
│       ├── similarity/    # Similarity calculation algorithms
│       └── text/          # Text processing utilities
├── tests/                 # Test files
│   ├── integration/       # Integration tests
│   └── unit/              # Unit tests
└── wrangler.jsonc         # Cloudflare Workers configuration

Technical Details

  • Built on Cloudflare Workers platform
  • Uses LRU caching for performance optimization
  • Implements weighted multi-algorithm similarity scoring
  • Supports both character and word-level matching for Chinese text
  • Configurable similarity thresholds and weights
  • Uses inverted index for faster preliminary search

Performance Optimization

  • Implements caching with LRU (Least Recently Used) strategy
  • Maximum cache size: 2000 entries
  • Minimum score threshold: 0.08
  • Optimized similarity calculations for both English and Chinese text
  • Two-tier search strategy: inverted index for fast preliminary results, followed by detailed scoring

Response Format

All endpoints return responses in the following format:

interface ResponseContent {
    type: 'text';
    text: string;
}

Development

This project is built using TypeScript and Cloudflare Workers. The main functionality is implemented in the RemixIconMCP class which extends WorkerEntrypoint.

Setup and Deployment

# Install dependencies
npm install

# Run development server
npm run dev

# Deploy to Cloudflare Workers
npm run deploy

# Run tests
npm run test

License

MIT License

Tools

No tools

Comments

Recommend MCP Servers

View All MCP Servers