Stitch MCP Auto

🚀 Stitch MCP Auto 🎨

One command. Instant UI. The most automated MCP server for Google Stitch.

English | 한국어

One command. Instant UI. The most automated MCP server for Google Stitch.

💡 Just share this link with your AI: https://github.com/GreenSheep01201/stitch-mcp-auto

✨ Features

• Auto Setup - One command installs everything (gcloud auth, API enable, MCP config)
• Multi-CLI Support - Works with Claude Code, Gemini CLI, Codex CLI
• 19 Custom Tools + Stitch Core - Design generation, accessibility checks, tokens, responsive variants, and design system export
• 7 Workflow Commands - /design, /design-system, /design-flow, /design-qa, /design-export, /generate-asset, /design-full
• 🎨 AI Image Generation - Generate logos, icons, hero images via Gemini 3 Pro (uses Antigravity - Google's experimental image generation API)
• 🎭 Orchestration Mode - One prompt to generate assets + complete UI design
• 🌐 i18n Support - Auto-detects system language (English/Korean) for setup wizard and console messages

🚀 Quick Start

After installing Node.js and gcloud CLI:

npx stitch-mcp-auto-setup

The setup wizard will:

1. Open a web-based setup page at http://localhost:51121
2. Guide you through Google login
3. Help you select or create a project
4. Enable the Stitch API
5. Save your configuration

Note: When starting the MCP server directly (npx stitch-mcp-auto or via your editor), the setup wizard runs on http://localhost:8086.

🤖 AI Agent Note (WSL Environment): If running in WSL, the browser cannot open automatically. Open the URL shown in the terminal (usually http://localhost:51121 for setup.js, or http://localhost:8086 when the MCP server starts) in your Windows browser:

# Run this in PowerShell (not WSL terminal)
Start-Process "http://localhost:51121"
# If you see 8086 instead:
Start-Process "http://localhost:8086"

Or instruct the user to manually open the shown URL in their Windows browser.

📦 Installation

Prerequisites (Install These First)

1. Install Node.js (v18 or higher)

Windows

# Using winget (Windows 10/11)
winget install OpenJS.NodeJS.LTS

# Or download installer from: https://nodejs.org/

macOS

# Using Homebrew
brew install node@22

# Or download installer from: https://nodejs.org/

Linux / WSL

# Using nvm (recommended)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
source ~/.bashrc
nvm install 22
nvm use 22

# Or using apt (Ubuntu/Debian)
curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash -
sudo apt-get install -y nodejs

Verify Node.js installation:

node --version   # Should show v18.x.x or higher
npm --version    # Should show 9.x.x or higher

2. Install Google Cloud CLI (gcloud)

Windows

Option A: Installer (Recommended)

1. Download from Google Cloud SDK
2. Run the installer (check "Run gcloud init" option)
3. Restart your terminal

Option B: PowerShell

(New-Object Net.WebClient).DownloadFile("https://dl.google.com/dl/cloudsdk/channels/rapid/GoogleCloudSDKInstaller.exe", "$env:TEMP\gcloud.exe")
Start-Process "$env:TEMP\gcloud.exe" -Wait
# Restart terminal after installation

macOS

# Using Homebrew (Recommended)
brew install --cask google-cloud-sdk

# Or direct install
curl https://sdk.cloud.google.com | bash
exec -l $SHELL

Linux / WSL

# Download and install
curl -sSL https://sdk.cloud.google.com | bash -s -- --disable-prompts --install-dir=$HOME

# Add to PATH
echo 'export PATH="$HOME/google-cloud-sdk/bin:$PATH"' >> ~/.bashrc
source ~/.bashrc

Verify gcloud installation:

gcloud --version
# Should show: Google Cloud SDK xxx.x.x

AI Agent Quick Start (Semi-Automated)

This repo supports semi-automated setup with a few unavoidable manual steps (OAuth login).

Checklist (agent-friendly):

1. Run setup: npm run setup (or npx -p stitch-mcp-auto stitch-mcp-auto-setup)
2. Open the URL shown in terminal (usually http://localhost:51121)
3. Complete Google login in the browser (gcloud + Stitch API enable)
4. ⚠️ CRITICAL: Verify MCP registration (see below)
5. Start server: npm start or let the editor launch it

Manual-only steps:

• Browser OAuth consent (Google login)
• WSL browser opening (must open URL manually)

Environment-specific commands:

• Windows (PowerShell):
  • Open setup URL: Start-Process "http://localhost:51121" (or http://localhost:8086)
• macOS:
  • Open setup URL: open "http://localhost:51121" (or http://localhost:8086)
• Linux:
  • Open setup URL: xdg-open "http://localhost:51121" (or http://localhost:8086)
• WSL (Windows host):
  • Run in Windows PowerShell: Start-Process "http://localhost:51121"

⚠️ Post-Installation MCP Verification (REQUIRED)

After setup completes, always verify that the MCP server is registered:

# Claude Code
claude mcp list | grep stitch

# Gemini CLI
gemini mcp list | grep stitch

# Codex CLI
codex mcp list | grep stitch

If stitch is NOT listed, manually register:

Why verification is needed: The setup wizard writes directly to config files (~/.claude.json, ~/.gemini/settings.json, ~/.codex/config.toml), but verification ensures the configuration is correct. Always verify to avoid "MCP server not found" errors.

Other Verification (optional):

• node auth.js --status (shows token + project status)
• npm start (server boots and prints "Ready")

Manual Installation (Alternative)

Step 1: Run Auto Setup

npx -p stitch-mcp-auto stitch-mcp-auto-setup

Or if you cloned the repository:

node setup.js

Setup Process

1. Welcome Page - Click "Login with Google" button
2. Google Login - A new browser window opens for authentication
   • After login, you may see "Connection refused" page - just close it
   • The setup page will automatically detect your login
3. Project Selection - Choose an existing project or create a new one
4. API Activation - Click the button to enable Stitch API
   • Wait for activation (may take a few seconds)
5. Complete - Copy the MCP configuration to your editor

⚠️ WSL Users - IMPORTANT: WSL cannot open browsers directly. When the setup wizard starts:

1. Copy the URL http://localhost:51121
2. Open it manually in your Windows browser (Chrome, Edge, etc.)
3. Or run in PowerShell: Start-Process "http://localhost:51121"

The authentication will work because WSL shares localhost with Windows.

Step 3: Configure MCP Client

After setup completes, add the configuration to your MCP client.

Claude Desktop

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "stitch": {
      "command": "npx",
      "args": ["-y", "stitch-mcp-auto"],
      "env": {
        "GOOGLE_CLOUD_PROJECT": "YOUR_PROJECT_ID"
      }
    }
  }
}

Claude Code

🌐 Method 1: CLI Command (Recommended)

The setup wizard automatically writes to ~/.claude.json. To add manually via CLI:

claude mcp add -e GOOGLE_CLOUD_PROJECT=YOUR_PROJECT_ID -s user stitch -- npx -y stitch-mcp-auto

Or edit ~/.claude.json directly:

{
  "mcpServers": {
    "stitch": {
      "command": "npx",
      "args": ["-y", "stitch-mcp-auto"],
      "env": {
        "GOOGLE_CLOUD_PROJECT": "YOUR_PROJECT_ID"
      }
    }
  }
}

📁 Method 2: Per-project Settings

Create .mcp.json in your project root:

{
  "mcpServers": {
    "stitch": {
      "command": "npx",
      "args": ["-y", "stitch-mcp-auto"],
      "env": {
        "GOOGLE_CLOUD_PROJECT": "YOUR_PROJECT_ID"
      }
    }
  }
}

Gemini CLI

🌐 Method 1: CLI Command (Recommended)

gemini mcp add stitch -- npx -y stitch-mcp-auto --env GOOGLE_CLOUD_PROJECT=YOUR_PROJECT_ID

Or edit ~/.gemini/settings.json directly:

{
  "mcpServers": {
    "stitch": {
      "command": "npx",
      "args": ["-y", "stitch-mcp-auto"],
      "env": {
        "GOOGLE_CLOUD_PROJECT": "YOUR_PROJECT_ID"
      }
    }
  }
}

Codex CLI

🌐 Method 1: CLI Command (Recommended)

codex mcp add stitch -- npx -y stitch-mcp-auto --env GOOGLE_CLOUD_PROJECT=YOUR_PROJECT_ID

Or edit ~/.codex/config.toml directly:

[mcp_servers.stitch]
command = "npx"
args = ["-y", "stitch-mcp-auto"]

[mcp_servers.stitch.env]
GOOGLE_CLOUD_PROJECT = "YOUR_PROJECT_ID"

Cursor

Go to Settings > MCP > Add New Server and add:

• Command: npx
• Args: -y stitch-mcp-auto
• Environment: GOOGLE_CLOUD_PROJECT=YOUR_PROJECT_ID

Optional: GOOGLE_CLOUD_PROJECT is only used for Stitch image fallback and auth.js --status. The setup wizard stores the active project in ~/.stitch-mcp-auto/config.json, and workspace tools save .stitch-project.json in your current folder.

💻 Usage Examples

Create a New Project

Create a new Stitch project called "My App"

Generate a Screen

Generate a login page with email and password fields,
a "Forgot Password" link, and social login buttons for Google and Apple.
Use a modern gradient background.

Generate with Specific Style

Create a dashboard screen with:
- Dark theme
- Sidebar navigation
- 4 stat cards at the top
- A line chart showing weekly data
- A recent activity list

Multi-language Support

Create a Korean e-commerce product detail page with:
- Product image carousel
- Price and discount badge
- Size selector
- Add to cart button
- Customer reviews section

The "Designer Flow"

For consistent UI across multiple screens:

1. Extract context from existing screen:

   Get the design context from the Home Screen in project X
   

2. Generate new screen with same style:

   Using that design context, generate a Settings screen
   with the same visual style
   

Using Design Tokens

Export design tokens for your development workflow:

Generate CSS variables from the dashboard screen's design

Create a Tailwind config based on the home screen

Trending Design Generation

Apply modern UI/UX trends automatically:

Create a pricing page with glassmorphism and gradient-mesh effects

Design a dashboard using bento-grid layout and dark mode

Batch Screen Generation

Create multiple screens at once with consistent styling:

Generate a complete onboarding flow: welcome, features, signup, and confirmation screens

Accessibility Check

Ensure your designs are accessible:

Check the login page for WCAG AA compliance

Design System Export

Export for developer handoff:

Export the complete design system from this project including tokens and components

AI Image Generation (NEW in v1.1.0)

Generate design assets with AI:

/generate-asset logo "Eco-friendly organic food delivery service called GreenBite"

/generate-asset hero "Modern fintech app showing financial growth" --style gradient --ratio 16:9

/generate-asset icon "Shopping cart with checkmark" --style flat --colors "#4CAF50"

Full Design Orchestration (NEW in v1.1.0)

One prompt to complete design - automatically generates assets and creates UI:

/design-full "친환경 유기농 식품 쇼핑몰 메인 페이지. 녹색 테마, 신선한 느낌, 모던한 디자인"

/design-full "AI-powered project management tool landing page. Professional, blue gradient theme, with pricing section"

Note: The orchestration mode automatically:

1. Analyzes required assets (logo, hero, icons)
2. Generates each asset using Gemini 3 Pro (requires Antigravity auth)
3. Creates complete UI screen with Stitch API
4. Returns all assets + final UI in one response

Without Antigravity auth: UI screen is still generated, but without custom image assets.

📚 Documentation

Available Tools

🔄 Automatic Project Management (NEW)

No more manual projectId passing! The server automatically manages project context:

1. Auto-detection: When you call tools like generate_screen_from_text without projectId, it automatically uses the workspace project
2. Auto-save: When you create_project, it's automatically saved to .stitch-project.json in your current folder
3. Session continuity: Return to the same folder later, and your project is automatically loaded

How it works:

┌─────────────────────────────────────────────────────────────┐
│  Tool called without projectId                              │
│                      ↓                                      │
│  1. Check active session project                            │
│  2. If none → Load from .stitch-project.json                │
│  3. If none → Return "PROJECT_REQUIRED" with options        │
│                      ↓                                      │
│  User creates/selects project → Auto-saved to workspace     │
└─────────────────────────────────────────────────────────────┘

Manual workspace tools (optional):

Core Stitch Tools (API + convenience wrappers)

🎨 Professional Web Design Tools

Design Consistency Tools

Productivity Tools

Quality & Analysis Tools

Design Enhancement Tools

🎨 AI Image Generation Tools (v1.1.0)

These tools have different authentication requirements:

Supported Models for Image Generation

See also: ANTIGRAVITY-MODELS.md (EN) / ANTIGRAVITY-MODELS.ko.md (KO) for the full model list, thinking variants, and endpoint details.

💡 Model Selection: You can specify a model when calling generate_design_asset:

{
  "assetType": "logo",
  "prompt": "Modern tech company logo",
  "model": "gemini-3-pro"
}

Key parameters (generate_design_asset):

• assetType: logo, icon, illustration, hero, wireframe, background, pattern
• aspectRatio: 1:1, 16:9, 9:16, 4:3, 3:4
• saveToFile: true by default (saves to current directory)
• forceAntigravityAuth: true to trigger browser login

📋 Role Separation:

• Stitch API (gcloud auth): UI page/screen generation - available to all users
• Antigravity OAuth: Image asset generation (logo, icon, hero) - requires separate authentication

Stitch-only users can still create complete UI pages using generate_screen_from_text. Antigravity extends this with custom AI-generated assets.

🔲 Background Removal (NEW)

generate_design_asset supports automatic background removal for transparent assets:

Example:

{
  "assetType": "logo",
  "prompt": "Modern tech company logo with abstract shapes",
  "removeBackground": true,
  "backgroundRemovalMode": "white",
  "backgroundThreshold": 240
}

This produces a transparent PNG, ideal for:

• Logos - No white border when placed on colored backgrounds
• Icons - Clean integration into UI designs
• Illustrations - Seamless overlay on any background

MCP Prompts (Auto-Discovered)

MCP Prompts are automatically available when the server is connected - no installation required. These appear in your MCP client's prompt list.

Usage in Claude Code:

/stitch:design login page with dark mode
/stitch:design-system settings page
/stitch:design-flow onboarding: welcome -> signup -> complete
/stitch:design-qa all --level AA
/stitch:design-export --token_format tailwind

Note: MCP Prompts include detailed workflow instructions that guide the AI through complex design tasks automatically.

Custom Commands (Multi-CLI Support)

When you run npx -p stitch-mcp-auto stitch-mcp-auto-setup, the setup wizard automatically installs custom commands for all major AI CLI tools - Claude Code, Gemini CLI, and Codex CLI.

Supported CLIs

Available Commands

CLI Usage Examples

Claude Code:

/design login page dark mode
/design-system settings page
/design-flow onboarding: welcome -> signup -> complete

Gemini CLI:

/stitch:design login page dark mode
/stitch:design-system settings page
/stitch:design-flow onboarding: welcome -> signup -> complete

Codex CLI:

$stitch-design login page dark mode
$stitch-design-system settings page
$stitch-design-flow onboarding: welcome -> signup -> complete

Commands Installation Location

Commands are automatically installed to all CLI directories:

~/.claude/commands/          # Claude Code (Markdown)
├── design.md
├── design-system.md
├── design-flow.md
├── design-qa.md
├── design-export.md
├── generate-asset.md        # NEW in v1.1.0
└── design-full.md           # NEW in v1.1.0

~/.gemini/commands/stitch/   # Gemini CLI (TOML)
├── design.toml
├── design-system.toml
├── design-flow.toml
├── design-qa.toml
├── design-export.toml
├── generate-asset.toml      # NEW in v1.1.0
└── design-full.toml         # NEW in v1.1.0

~/.codex/skills/stitch/      # Codex CLI (Skills)
├── design.md
├── design-system.md
├── design-flow.md
├── design-qa.md
├── design-export.md
├── generate-asset.md        # NEW in v1.1.0
└── design-full.md           # NEW in v1.1.0

🔧 Technical Details

Troubleshooting

"gcloud: command not found"

Linux/macOS/WSL:

export PATH="$HOME/google-cloud-sdk/bin:$PATH"
echo 'export PATH="$HOME/google-cloud-sdk/bin:$PATH"' >> ~/.bashrc

Windows: Restart your terminal after installing gcloud.

"Stitch API has not been used in project" Error

Enable the API manually:

gcloud services enable stitch.googleapis.com --project=YOUR_PROJECT_ID

Or visit: https://console.cloud.google.com/apis/library/stitch.googleapis.com?project=YOUR_PROJECT_ID

Token Expired / Authentication Error

Refresh your authentication:

gcloud auth login

Or re-run the setup:

npx -p stitch-mcp-auto stitch-mcp-auto-setup

"Connection Refused" After Google Login

This is normal behavior. After Google authentication completes, the browser redirects to localhost:8085 which is a temporary callback server that gcloud uses. Once authentication is complete, this server closes.

Solution: Simply close this tab and return to the setup page. It will automatically detect your login.

Browser Doesn't Open Automatically (WSL)

If the browser doesn't open automatically in WSL:

1. Copy the URL shown in the terminal
2. Paste it in your Windows browser manually
3. Complete the login
4. Return to the setup page

Full Reset

If something goes wrong, reset everything:

# Remove stitch-mcp-auto config
rm -rf ~/.stitch-mcp-auto

# Revoke gcloud credentials
gcloud auth revoke --all

# Re-run setup
npx -p stitch-mcp-auto stitch-mcp-auto-setup

Architecture

┌──────────────────────────────────────────────────────────────┐
│                        User Request                          │
│              "Create a login page with..."                   │
│              "/design-full eco-friendly shop"                │
└──────────────────────────────────────────────────────────────┘
                              │
                              ▼
┌──────────────────────────────────────────────────────────────┐
│                    AI Editor (Claude/Cursor)                 │
│                         MCP Client                           │
└──────────────────────────────────────────────────────────────┘
                              │
                              ▼
┌──────────────────────────────────────────────────────────────┐
│                     stitch-mcp-auto                          │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐           │
│  │ setup.js    │  │ index.js    │  │ auth.js     │           │
│  │ (Auto Setup)│  │ (MCP Server)│  │ (OAuth)     │           │
│  └─────────────┘  └─────────────┘  └─────────────┘           │
│                          │                                   │
│           ┌──────────────┼──────────────┐                    │
│           ▼              ▼              ▼                    │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐           │
│  │ Design Tools│  │ Image Gen   │  │ Orchestrate │           │
│  │ (UI + QA)   │  │ (Antigrav.) │  │ (assets+UI) │           │
│  └─────────────┘  └─────────────┘  └─────────────┘           │
└──────────────────────────────────────────────────────────────┘
                              │
            ┌─────────────────┼─────────────────┐
            ▼                                   ▼
┌─────────────────────────────┐   ┌─────────────────────────────┐
│   Google Cloud Platform     │   │   Antigravity (Optional)    │
│  ┌─────────┐  ┌─────────┐   │   │  ┌─────────────────────┐    │
│  │ gcloud  │  │ Stitch  │   │   │  │ Gemini 3 Pro        │    │
│  │ CLI     │  │ API     │   │   │  │ (Image Generation)  │    │
│  │ (Auth)  │  │ (UI Gen)│   │   │  │ FREE via OAuth      │    │
│  └─────────┘  └─────────┘   │   │  └─────────────────────┘    │
└─────────────────────────────┘   └─────────────────────────────┘

📋 Role Separation:

• Stitch API (gcloud): Handles all UI page/screen generation
• Antigravity OAuth: Handles image asset generation (logo, icon, hero images)

Without Antigravity authentication, generate_design_asset will return an error. Use Stitch's generate_screen_from_text for UI pages.

File Structure

stitch-mcp-auto/
├── index.js               # Main MCP server (with Antigravity OAuth)
├── setup.js               # Web-based auto setup wizard
├── auth.js                # OAuth helper utilities
├── package.json           # Dependencies and scripts
├── README.md              # Documentation (EN)
├── README.ko.md           # Documentation (KO)
├── ANTIGRAVITY-MODELS.md  # Antigravity model list (EN)
├── ANTIGRAVITY-MODELS.ko.md # Antigravity model list (KO)
├── AGENTS.md              # Repository guidelines
└── skills/                # Command sources (Claude/Gemini/Codex)
    ├── design.md
    ├── design-system.md
    ├── design-flow.md
    ├── design-qa.md
    ├── design-export.md
    ├── generate-asset.md    # NEW: AI image generation
    └── design-full.md       # NEW: Full orchestration

Configuration Files

Scripts

📄 License

Apache 2.0 - Open source and free to use.

Credits

• Author: GreenSheep01201 (Wongil Seo)
• Based on: stitch-mcp by Kargatharaakash
• Concept: Automated implementation of the MCP (Model Context Protocol) file stitching system.

Support

• Issues: GitHub Issues
• Documentation: Google Stitch API

Built with love for the AI community