mini-gamma-slides

gkdivya/mini-gamma-slides

3.2

If you are the rightful owner of mini-gamma-slides and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

A powerful MCP server that enables AI agents to create professional presentations using Google Slides templates.

Tools
5
Resources
0
Prompts
0

Mini Gamma-like Presentation Generator

A powerful MCP (Model Context Protocol) server that enables AI agents to create professional presentations using Google Slides templates. This system provides Gamma-like functionality for automated presentation generation with template-based design consistency.

🚀 Features

Core Capabilities

  • Template Management: Registry-based template system for different presentation types
  • Template Cloning: Automatic template duplication via Google Drive API
  • Layout Extraction: Analyze template structure for AI content planning
  • Content Injection: Replace placeholders with generated content
  • Multi-Category Support: Business, marketing, startup, academic templates

AI Integration Ready

  • Layout-Aware Content: AI can understand template structure before generating content
  • Placeholder Mapping: Structured content injection system
  • Agent-Friendly: Designed for seamless AI agent integration
  • Batch Operations: Efficient multi-slide creation and editing

📋 Available Templates

Business Templates

  • Professional Business: Clean corporate presentations, pitch decks
  • Startup Pitch: Investor-ready deck with standard startup sections

Marketing Templates

  • Marketing Campaign: Vibrant, engaging marketing presentations
  • Product Launch: Feature-focused presentation layouts

Academic Templates

  • Minimal Clean: Academic and technical presentations
  • Research: Research paper and thesis presentations

🛠️ Setup

Prerequisites

  • Python 3.8+
  • uv - Fast Python package installer
  • Google Cloud Project with APIs enabled:
    • Google Slides API
    • Google Drive API
  • OAuth2 credentials (client ID, client secret, refresh token)

Quick Setup with uv

  1. Install uv (if not already installed):
curl -LsSf https://astral.sh/uv/install.sh | sh
  1. Clone and setup environment:
cd /path/to/your/workspace/Google_MCP
./setup.sh

This will:

  • Create a virtual environment with uv
  • Install all required dependencies
  • Create a .env template file
  1. Configure Google API credentials: Edit the .env file created by setup:
GOOGLE_CLIENT_ID=your_client_id_here
GOOGLE_CLIENT_SECRET=your_client_secret_here  
GOOGLE_REFRESH_TOKEN=your_refresh_token_here
  1. Set up templates:
  • Copy your template presentations to Google Drive
  • Update templates.json with your template file IDs
  • Ensure templates have appropriate placeholders ({{TITLE}}, {{CONTENT}}, etc.)
  1. Start the MCP server:
./run.sh

Server will start on http://127.0.0.1:8101

Development Setup

For development with additional tools (pytest, black, flake8, mypy):

./setup-dev.sh

Manual Installation (Alternative)

If you prefer manual setup:

# Create virtual environment
uv venv

# Activate virtual environment  
source .venv/bin/activate

# Install dependencies
uv pip install mcp google-api-python-client google-auth google-auth-oauthlib google-auth-httplib2 python-dotenv uvicorn fastapi pydantic requests aiohttp

# Run server
python google_slides_mcp_sse.py

🎯 Usage

MCP Tools Available

Template Management
  • get_templates() - List available templates
  • get_templates(category="business") - Filter by category
  • extract_layouts(template_name="business") - Analyze template structure
Presentation Creation
  • clone_template(template_name, new_title) - Clone template
  • create_presentation_from_template(template_name, new_title, content_data) - Full workflow
Standard Slides Operations
  • create_presentation(title) - Create blank presentation
  • get_presentation(presentationId) - Get presentation details
  • batch_update_presentation(presentationId, requests) - Batch updates
  • summarize_presentation(presentationId) - Extract content

Example Workflows

1. Business Presentation Creation
# Get business templates
templates = await get_templates(category="business")

# Analyze template structure
layouts = await extract_layouts(template_name="business")

# Create presentation with content
result = await create_presentation_from_template(
    template_name="business",
    new_title="Q4 2024 Business Review",
    content_data={
        "company_name": "TechCorp Industries",
        "presenter": "Sarah Johnson, CEO",
        "title": "Strategic Business Review",
        "content": "Revenue growth of 127% year-over-year..."
    }
)
2. Marketing Campaign Deck
# Create marketing presentation
result = await create_presentation_from_template(
    template_name="marketing", 
    new_title="Green Revolution Campaign 2024",
    content_data={
        "brand_name": "EcoLife Products",
        "campaign_title": "Green Revolution 2024", 
        "value_proposition": "Sustainable living made simple",
        "cta": "Join the Green Revolution Today!"
    }
)
3. AI Agent Integration
# Agent workflow
async def create_ai_presentation(user_request: str):
    # 1. Analyze request and determine template type
    template_type = analyze_request_type(user_request)
    
    # 2. Get template structure for content planning
    layouts = await extract_layouts(template_name=template_type)
    
    # 3. Generate content using GPT with layout awareness
    content_data = await generate_content_for_template(
        user_request, layouts, template_type
    )
    
    # 4. Create presentation
    return await create_presentation_from_template(
        template_name=template_type,
        new_title=extract_title(user_request),
        content_data=content_data
    )

🏗️ Template Structure

Template Registry Format

{
  "template_name": {
    "fileId": "google_drive_file_id",
    "name": "Human Readable Name",
    "description": "Template description", 
    "category": "business|marketing|academic",
    "layouts": {
      "slide_type": "GOOGLE_LAYOUT_TYPE"
    },
    "placeholders": {
      "key": "{{PLACEHOLDER_PATTERN}}"
    }
  }
}

Placeholder System

Templates use placeholder patterns that get replaced with actual content:

  • {{COMPANY_NAME}} → Company name
  • {{TITLE}} → Slide title
  • {{CONTENT}} → Main content
  • {{DATE}} → Presentation date
  • {{PRESENTER}} → Presenter name

🤖 AI Agent Integration

Content Planning with Layout Awareness

The extract_layouts tool provides structure information that AI agents can use to:

  • Understand available slide layouts
  • Plan content appropriate for each layout type
  • Map generated content to correct placeholders
  • Ensure content fits layout constraints

Example Agent Prompt

You are creating a marketing presentation using the "marketing" template.

Available layouts: hero, feature, stats, testimonial, cta
Template placeholders: brand_name, campaign_title, value_proposition, features, testimonial, cta

Generate content that fits these layouts and maps to the placeholders appropriately.

📊 Template Layout Types

Standard Google Slides Layouts

  • TITLE_SLIDE - Title and subtitle
  • TITLE_AND_BODY - Title with body content
  • TITLE_AND_TWO_COLUMNS - Title with two-column layout
  • SECTION_HEADER - Section divider slide
  • BLANK - Empty slide

Custom Template Layouts

Templates can define semantic layout names:

  • hero - Main presentation slide
  • problem - Problem statement
  • solution - Solution overview
  • features - Feature highlights
  • testimonial - Customer testimonials
  • cta - Call to action

🔧 Advanced Features

Batch Slide Creation

Create multiple slides with different layouts in one operation:

requests = [
    {"createSlide": {"slideLayoutReference": {"predefinedLayout": "TITLE_SLIDE"}}},
    {"createSlide": {"slideLayoutReference": {"predefinedLayout": "TITLE_AND_BODY"}}},
    # Add content to slides
    {"insertText": {"objectId": "slide1_title", "text": "My Title"}},
]

await batch_update_presentation(presentation_id, requests)

Custom Content Injection

Beyond simple placeholder replacement:

  • Insert images and shapes
  • Apply custom formatting
  • Add charts and tables
  • Set slide transitions

🚦 Error Handling

The system includes comprehensive error handling for:

  • Invalid template names
  • Missing Google API credentials
  • Drive API permission issues
  • Malformed content data
  • Template cloning failures

🔐 Security & Permissions

Required Google OAuth Scopes

  • https://www.googleapis.com/auth/presentations - Slides API access
  • https://www.googleapis.com/auth/drive - Drive API for template cloning

Template Access

  • Templates must be accessible by the service account/OAuth user
  • Cloned presentations inherit the same permissions
  • Consider using shared drives for team templates

📈 Performance Considerations

  • Template registry is loaded once at startup
  • Layout extraction is cached per template
  • Batch operations are preferred for multiple slides
  • Large templates may have slower clone operations

🧪 Testing

Run the example usage script:

python example_usage.py

This demonstrates the complete workflow and shows expected output formats.

🔄 Future Enhancements

  • Dynamic template creation from existing presentations
  • Template marketplace integration
  • Advanced content generation with layout optimization
  • Multi-language template support
  • Custom theme and branding injection
  • Integration with design systems
  • Real-time collaboration features

🤝 Contributing

  1. Add new template types to templates.json
  2. Extend placeholder system for new content types
  3. Add new MCP tools for advanced functionality
  4. Improve AI agent integration patterns

📝 License

This project is designed for educational and development purposes. Ensure you have appropriate Google API quotas and permissions for production use.

Ready to build the future of AI-powered presentation creation! 🎨✨