Stitch Design Agent
Brings Google Stitch design tools into AI assistants, enabling browsing projects, viewing screen designs, extracting design tokens, and generating new screens via natural language.
README
Stitch Design Agent - MCP App
An interactive MCP App that brings Google Stitch design tools directly into AI assistants like Claude Code, Claude Desktop, and ChatGPT. Browse projects, view screen designs, extract design tokens, and generate new screens — all within your chat interface.
Built with the MCP Apps (ext-apps) extension, which renders interactive React UI inside MCP host sandboxed iframes.
Screenshots
Browse all your Stitch design projects with thumbnail previews
View all screens within a project with device type badges
Split panel view: screen preview with HTML/CSS code and design tokens
Features
| Tool | Description |
|---|---|
list-projects |
Browse all Stitch design projects with thumbnail previews, device types, and theme colors |
list-screens |
View all screens within a project as a thumbnail grid |
design-viewer |
Split panel showing screen preview alongside HTML/CSS code and extracted design tokens |
generate-design |
Generate new screen designs from text prompts using AI, with follow-up suggestions |
extract-design-context |
Extract design tokens (colors, fonts, spacing, layouts) for brand consistency |
Installation
Prerequisites
- Node.js >= 18
- A Google Cloud project with the Stitch API enabled
- Application Default Credentials configured
# Authenticate with Google Cloud
gcloud auth application-default login
Setup
# Clone the repo
git clone https://github.com/jayeshvpatil/stitch-mcp-app.git
cd stitch-mcp-app
# Install dependencies
npm install
# Build
npm run build
Configure your MCP client
Add to your MCP client configuration (e.g. ~/.claude/mcp.json for Claude Code):
{
"mcpServers": {
"stitch-design": {
"command": "node",
"args": ["/path/to/stitch-mcp-app/dist/main.js"],
"env": {
"GOOGLE_CLOUD_PROJECT": "your-gcp-project-id"
}
}
}
}
Replace /path/to/stitch-mcp-app with the actual path and your-gcp-project-id with your Google Cloud project ID.
Usage
Once configured, you can interact with your Stitch designs through natural language:
"Show me my Stitch design projects"
"List screens in the Sentiment Dashboard project"
"View the Data Explorer screen design"
"Generate a mobile login screen with dark theme"
"Extract design tokens from this screen"
The app renders interactive UI cards directly in the chat — click project cards to drill into screens, click screens to view full designs with code.
Architecture
stitch-mcp-app/
├── server.ts # MCP server: 5 tools + 1 shared UI resource
├── main.ts # stdio transport entry point
├── stitch-client.ts # Stitch REST API client (google-auth-library)
├── mcp-app.html # HTML entry for React UI
├── vite.config.ts # Vite + vite-plugin-singlefile
└── src/
├── StitchApp.tsx # Main component: useApp() hook + view routing
├── types.ts # TypeScript interfaces matching Stitch API
├── global.css # Dark mode responsive styles
└── components/
├── ProjectList.tsx # Project card grid
├── ScreenList.tsx # Screen thumbnail grid
├── DesignViewer.tsx # Split panel (image + code/tokens)
├── GenerateDesign.tsx # AI generation + suggestions
└── DesignContext.tsx # Design token display
Key patterns:
- Single shared UI resource (
ui://stitch-design/app.html) across all 5 tools - React UI bundled into a single HTML file via Vite +
vite-plugin-singlefile - Tools return
{ type, data }JSON — React app routes views viaontoolresult - Direct Stitch API calls using
google-auth-library(not proxied)
Development
# Watch mode (rebuilds on changes)
npm run dev
# Build UI only
npm run build:ui
# Build server only
npm run build:server
# Full build
npm run build
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.