mcpField
L
lucide-icons-mcp

lucide-icons-mcp

Local
github.com/SeeYangZhiDeveloper Tools43Updated 16w ago

About lucide-icons-mcp

lucide-icons-mcp is an MCP server that exposes the Lucide icon library as resources and tools for LLMs and agent applications, built with Bun and the MCP TypeScript SDK. It suits frontend developers who want coding assistants to find and use icons.

Key features

  1. 01Search over 1,500 Lucide icons by name or category
  2. 02Browse icons by category
  3. 03Get React/JSX usage examples for an icon
  4. 04Run in either HTTP or stdio mode
  5. 05Test locally with the MCP Inspector

Use cases

  1. 01Finding a suitable icon and generating JSX in a coding assistant
  2. 02Browsing available icons by design category
  3. 03Providing icon references while building a UI in Claude Desktop

MCP support

The server exposes icon search, category browsing and usage example tools to MCP clients so AI coding assistants can call them.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

No install method listed yet

We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.

Repository

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

lucide-icons-mcp

Ask DeepWiki

README image

A Model Context Protocol (MCP) server exposing Lucide React icons as resources and tools for LLMs and agentic applications. Built with Bun and the MCP TypeScript SDK.

What is Lucide?

Lucide is a beautiful & consistent icon toolkit made by the community. It's an open-source icon library that provides over 1,500+ carefully crafted icons in a single, consistent style. Originally forked from Feather Icons, Lucide has grown into one of the most popular icon libraries for modern web development.

What is MCP?

The Model Context Protocol (MCP) is a standard for AI tools to request specific context from sources outside their main training data.

This MCP server allows AI coding assistants and other agentic applications to access information about Lucide React icons, enabling better assistance with icon search, discovery, and implementation.

Features

  • 🔍 Icon Search: Search through 1,500+ Lucide icons by name or category
  • 📂 Category Browsing: List icons by categories (Design, Communication, Media, etc.)
  • 💡 Usage Examples: Get React/JSX code examples for any icon
  • 🔧 Icon Information: Detailed information about each icon
  • 🚀 MCP Integration: Ready for Claude Desktop and other MCP clients
  • 🌐 Dual Mode: HTTP server or stdio-based MCP server
  • 📊 Comprehensive Coverage: All Lucide icons with proper JSX usage

Prerequisites

Getting Started (Development)

1. Clone the repository
git clone https://github.com/SeeYangZhi/lucide-icons-mcp.git
cd lucide-icons-mcp
2. Install Bun (if you don't have it)

Refer to the official Bun installation guide.
After installation, restart your terminal and check:

bun --version
3. Install dependencies
bun install
4. Build the project

This compiles the TypeScript source to JavaScript in the build directory.

bun run build

Usage

HTTP Mode

You can run the HTTP server using npx:

npx lucide-icons-mcp

This starts the HTTP server (defaults to port 3000).

Or install globally:

npm install -g lucide-icons-mcp

Then run:

lucide-icons-mcp
Stdio Mode
npx lucide-icons-mcp --stdio
# or if installed globally
lucide-icons-mcp --stdio

Local Development

There are two main ways to run the MCP server:

1. HTTP Mode

Suitable for clients that support communication over HTTP.

For development (using Bun):

bun run start
# or directly
bun run src/entry.ts
2. Stdio Mode

Often used for direct integration with tools like Claude Desktop or the MCP Inspector.

For development (using Bun):

bun run src/entry.ts --stdio

Configuration with AI Tools

Example: Claude Desktop

To use this MCP server in Claude Desktop:

  1. Open your Claude Desktop configuration file:
code ~/Library/Application\ Support/Claude/claude_desktop_config.json
  1. Add the server to the mcpServers section:

Option A: via npx (Recommended):

{
  "mcpServers": {
    "lucide-icons": {
      "command": "npx",
      "args": ["lucide-icons-mcp", "--stdio"]
    }
  }
}

Option B: Pointing directly to the build output:

{
  "mcpServers": {
    "lucide-icons": {
      "command": "node",
      "args": ["/ABSOLUTE/PATH/TO/lucide-icons-mcp/build/entry.js", "--stdio"]
    }
  }
}
  1. Save the file and restart Claude Desktop.
  2. You should now see the "lucide-icons" server available in Claude's tools panel.

Tools Available (MCP)

This MCP server exposes the following tools to AI coding assistants. Each icon can belong to multiple categories, providing flexible search and discovery capabilities.

1. search_icons
  • Description: Search for Lucide icons by name or category. Returns icons with their complete category arrays.
  • Parameters:
    • query (string): Search term for icon name or category
    • category (optional string): Filter by specific category
    • limit (optional number): Maximum results to return
  • Returns: Array of icons with name and categories (string array)
2. search_categories
  • Description: Search for icon categories by category name using partial matching
  • Parameters:
    • query (string): Search term for category name
    • limit (optional number): Maximum results to return
  • Returns: Array of categories with name and iconCount
3. fuzzy_search_icons
  • Description: Fuzzy Search for icons from lucide by icon name
  • Parameters:
    • query (string): Search term for icon
    • limit (optional number): Maximum results to return
  • Returns: Array of icons with name and categories (string array)
4. fuzzy_search_categories
  • Description: Fuzzy Search for icon categories by category name
  • Parameters:
    • query (string): Search term for category name
    • limit (optional number): Maximum results to return
  • Returns: Array of categories with name and iconCount (string array)
5. get_icon_usage_examples
  • Description: Get usage examples for a Lucide React icon by the exact name
  • Parameters:
    • name (string): Search term for icon
  • Returns: React/JSX usage examples for a specific Lucide icon
6. list_all_icons_by_category
  • Description: List all icons in a specific category
  • Parameters:
    • category (string): Category name to list icons for
    • limit (optional number): Maximum results to return
7. list_all_categories
  • Description: List all available icon categories with their icon counts

Example Usage

Here's how an AI tool might use this MCP server:

Example 1: Finding Icons

User: "Find me icons related to ArrowRight"

AI tool calls search_icons:

{
  "query": "ArrowRight",
  "limit": 5
}

Response: Lists icons like ArrowRight

Example 2: Getting Usage Examples

User: "Show me how to use the ArrowRight icon"

AI tool calls get_icon_usage_examples:

{
  "name": "ArrowRight"
}

Response:

import { ArrowRight } from "lucide-react";

function Example() {
  return (
    <div>
      <ArrowRight />
    </div>
  );
}
Example 3: Browsing Categories

User: "What design-related icons are available?"

AI tool calls search_categories:

{
  "query": "design"
}

Then calls list_icons_by_category":

{
  "category": "Design",
  "limit": 10
}

Icon Categories

Lucide icons are organized into categories such as:

  • Accessibility
  • Accounts & access
  • Animals
  • Arrows
  • Brands
  • Buildings
  • Charts
  • Communication
  • Connectivity
  • Cursors
  • Design
  • Coding & development
  • Devices
  • Emoji
  • File icons
  • Finance
  • Food & beverage
  • Gaming
  • Home
  • Layout
  • Mail
  • Mathematics
  • Medical
  • Multimedia
  • Nature
  • Navigation
  • Notification
  • People
  • Photography
  • Science
  • Seasons
  • Security
  • Shapes
  • Shopping
  • Social
  • Sports
  • Sustainability
  • Text formatting
  • Time & calendar
  • Tools
  • Transportation
  • Travel
  • Weather

Testing MCP Locally with Inspector

You can test the MCP server locally using the MCP Inspector.

First, ensure the project is built:

bun run build

Then launch the Inspector:

npx @modelcontextprotocol/inspector node ./build/entry.js --stdio

This opens the Inspector interface for interactive testing of your MCP server.

Development Scripts

  • bun run dev: Starts the server in HTTP mode for development
  • bun run dev:stdio: Starts the stdio MCP server for development
  • bun run build: Compiles TypeScript to JavaScript (output in build/)
  • bun run lint: Lints the codebase using ESLint
  • bun run lint:fix: Automatically fixes linting issues
  • bun run crawl: Crawls Lucide website to update icon data
  • bun run pre-build: Crawls data, builds icon metadata, and fixes linting

Data Sources

This project includes a web crawler that automatically extracts icon information from the Lucide website, ensuring up-to-date icon data and categories.

Resources

License

This project is licensed under the MIT License. See LICENSE for details.

This project references icons from Lucide (https://lucide.dev/), which are licensed under their ISC License:

ISC License

Copyright (c) for portions of Lucide are held by Cole Bemis 2013-2022 as part of Feather (MIT). All other copyright (c) for Lucide are held by Lucide Contributors 2022.

Permission to use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.

THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.