@manansiingh/figma-react-mcp-server
Converts Figma designs into production-ready React components with design token extraction, widget registry integration, and micro-frontend module generation.
README
@manansiingh/figma-react-mcp-server
Enterprise-grade Figma-to-React MCP Server — converts Figma designs into production-ready React components with design token extraction, widget registry integration, and micro-frontend (MFE) module generation.
Why This Server?
There are several Figma MCP servers out there. Here's what makes this one different:
| Feature | Other Figma MCPs | This Server |
|---|---|---|
| Basic component generation | ✅ | ✅ |
| Design token extraction (colors, spacing, typography → reusable token file) | ❌ | ✅ |
| Registry-driven output (components self-register into a widget registry) | ❌ | ✅ |
| MFE-aware generation (independently deployable micro-frontend modules) | ❌ | ✅ |
| Auto Layout → Flexbox mapping | Partial | ✅ Full |
This server was built from real-world experience building enterprise design systems and micro-frontend dashboards at scale, not just weekend prototypes.
What It Does
1. Design Token Extraction
Instead of hardcoding colors and spacing into individual components, this server extracts Figma's design tokens into a standalone tokens.ts file:
// Auto-generated design tokens from Figma
export const tokens = {
colors: {
primary: '#6366F1',
background: '#F8FAFC',
text: '#0F172A',
},
spacing: {
sm: '8px',
md: '16px',
lg: '24px',
},
typography: {
heading: { fontFamily: 'Inter', fontSize: '24px', fontWeight: 700 },
body: { fontFamily: 'Inter', fontSize: '16px', fontWeight: 400 },
},
};
2. Widget Registry Output
Generated components automatically export a registry entry, making them plug-and-play in dashboard frameworks:
// Auto-generated registry entry
export const widgetManifest = {
id: 'figma-card-widget',
name: 'Card Widget',
component: CardWidget,
category: 'display',
version: '1.0.0',
};
3. Micro-Frontend Module Generation
Components are generated as independently deployable MFE modules with their own entry point, not just standalone .tsx files:
// Auto-generated MFE bootstrap
import { CardWidget } from './CardWidget';
import { tokens } from './tokens';
export default {
mount: (container: HTMLElement) => {
// Render into the provided container
},
unmount: (container: HTMLElement) => {
// Cleanup
},
};
Quick Start
Requirements
- A Figma Personal Access Token (Settings → Personal Access Tokens)
- A Figma File Key (from the URL:
figma.com/file/<FILE_KEY>/...) - A Node ID (from the URL when selecting a frame:
?node-id=1:2)
Add to Your MCP Client
Add this to your MCP client config (Claude Desktop, Cursor, Windsurf, etc.):
{
"mcpServers": {
"figma-to-react": {
"command": "npx",
"args": ["-y", "@manansiingh/figma-react-mcp-server"]
}
}
}
Usage
Once configured, simply prompt your AI agent:
"Use the figma-to-react tool to convert Node '123:4' in file 'abcxyz' into a React component with design tokens extracted. My Figma token is 'figd_XXX'."
Available MCP Tools
| Tool | Description |
|---|---|
generate_react_component |
Converts a Figma node into a production React component with inline styles derived from Auto Layout, fills, strokes, typography, and border radius. |
extract_design_tokens |
Pulls all color, spacing, and typography tokens from a Figma file into a reusable tokens.ts format. |
generate_mfe_module |
Generates a self-contained micro-frontend module with mount/unmount lifecycle hooks. |
How It Works
- Fetches the exact node tree from the Figma REST API
- Recursively parses geometry, auto-layout constraints, fills, strokes, and text styles
- Maps Figma's
layoutMode→ CSS Flexbox (display: flex,flexDirection,justifyContent,alignItems,gap) - Extracts design tokens into a separate reusable file
- Wraps output in registry-compatible and MFE-compatible formats
- Returns clean, production-ready React JSX via MCP tool response
Tech Stack
- Model Context Protocol SDK (stdio transport)
- Figma REST API
- TypeScript, Node.js (ES2022)
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。