slidecraft-mcp 0.1.0

SlideCraft MCP - Lightweight presentation generator with LLM-powered content creation

SlideCraft MCP 🚀

A lightweight Model Context Protocol server for generating rich presentations using LLM-powered content generation instead of heavyweight libraries.

🛠️ Available Tools

🚀 LLM-Powered Generation

• create_presentation_from_prompt: Generate complete presentations from single prompts
• generate_content_llm: Create slide content using AI (titles, bullets, summaries)
• create_chart_llm: Generate ASCII charts with descriptions

📝 Core Presentation Tools

• create_presentation: Create new lightweight presentations
• add_slide: Add slides with layouts
• add_text: Insert text content with positioning
• save_presentation: Save as lightweight JSON format

🎨 Visual Content (ASCII-Based)

• add_infographic: Create ASCII infographics (process, comparison, timeline, hierarchy)
• create_chart_llm: Generate bar, line, pie, scatter charts as ASCII art
• export_to_html: Export as HTML slideshow with themes
• export_to_markdown: Export as Markdown for documentation

📊 Export & Management

• list_presentations: View all active presentations
• export_to_html: Generate themed HTML slideshows
• export_to_markdown: Create Markdown documentation

Overview

This MCP server is designed for speed, efficiency, and AI-first content creation. Instead of relying on heavy dependencies like python-pptx or matplotlib, it uses:

• 🤖 LLM Integration: Generate complete presentations from prompts
• 🪶 Zero Heavy Dependencies: Pure Python with minimal requirements
• ⚡ Lightning Fast: Startup in milliseconds, not seconds
• 📊 ASCII Visualizations: Charts and infographics without plotting libraries
• � Lightweight Storage: JSON-based data structures
• 🌐 Multiple Exports: Markdown, HTML, and structured JSON

✨ Key Features

🪶 Ultra-Lightweight Architecture

• Zero heavyweight dependencies - No python-pptx, matplotlib, or PIL
• Fast startup - Loads in milliseconds
• Minimal memory footprint - Uses <50MB RAM
• JSON-based storage - Lightweight data structures

🤖 LLM-Powered Content Generation

• Complete presentation generation from single prompts
• Intelligent content creation - titles, bullets, summaries
• ASCII-based charts and infographics - no heavy plotting libraries
• Template-based enhanced output with emojis and structure

📊 Rich Media Support (Lightweight)

• ASCII art charts - bar, line, pie, scatter
• ASCII infographics - process flows, comparisons, timelines, hierarchies
• Multiple export formats - JSON, Markdown, HTML
• Themed outputs - default, dark, minimal, corporate

⚡ Performance Benefits

• Instant startup vs 2-3 seconds for heavy alternatives
• 10x smaller memory usage compared to traditional solutions
• No binary dependencies - pure Python implementation
• Cross-platform compatibility without external tools

Quick Start

Installation

1. Clone or download this repository
2. Install dependencies:

   uv sync --dev --all-extras
   

Usage

Claude Desktop Configuration

Add to your Claude Desktop configuration file:

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

{
  "mcpServers": {
    "slidecraft-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/slidecraft-mcp",
        "run",
        "slidecraft-mcp"
      ]
    }
  }
}

Direct Usage

# Run the server
uv run slidecraft-mcp

# For development
uv sync --dev
uv run slidecraft-mcp

🚀 Quick Usage Examples

Generate Complete Presentation from Prompt

# Create entire presentation with LLM
create_presentation_from_prompt({
  "prompt": "Benefits of microservices architecture",
  "slide_count": 5,
  "template": "business"
})

Add ASCII Infographic

# Add process flow infographic
add_infographic({
  "presentation_id": "abc123",
  "slide_index": 0,
  "infographic_type": "process",
  "data": "Development pipeline: Code → Build → Test → Deploy"
})

Export to Multiple Formats

# Export as HTML slideshow
export_to_html({
  "presentation_id": "abc123",
  "filename": "my_presentation.html",
  "theme": "corporate"
})

# Export as Markdown documentation  
export_to_markdown({
  "presentation_id": "abc123",
  "filename": "presentation.md"
})