Deckbuilder MCP Server

# Content-First Tools This directory contains the content-first analysis tools that were moved from the main MCP server to keep the core server focused on presentation generation. ## Files ### `content_first_tools.py` Contains the content-first analysis tools that implement the intelligent presentation consultation workflow: 1. **`analyze_presentation_needs_tool()`** - Analyzes user's presentation needs and recommends structure 2. **`recommend_slide_approach_tool()`** - Recommends optimal slide layouts based on content and intent 3. **`optimize_content_for_layout_tool()`** - Optimizes content structure and generates ready-to-use YAML ### `main.py` The streamlined MCP server containing only the core presentation generation tools: 1. **`create_presentation()`** - Create presentation from JSON data 2. **`create_presentation_from_file()`** - Create presentation from JSON or markdown files 3. **`create_presentation_from_markdown()`** - Create presentation from markdown content ## Why Were They Moved? The content-first tools were moved to separate the core presentation generation functionality from the intelligent content analysis features. This provides: - **Cleaner MCP server**: Focus on essential presentation generation - **Preserved functionality**: Content-first tools available for future integration - **Better maintainability**: Clear separation of concerns - **Reduced complexity**: Simpler MCP server interface ## Future Integration The content-first tools can be easily re-integrated into the MCP server by: 1. Adding the `@mcp.tool()` decorator back to each function 2. Adding the `ctx: Context` parameter back to each function signature 3. Importing the functions in `main.py` ## Content-First Workflow The preserved workflow implements a content-first design philosophy: ``` 1. analyze_presentation_needs_tool() ↓ (analyzes user intent and audience) 2. recommend_slide_approach_tool() ↓ (suggests optimal layouts with confidence scores) 3. optimize_content_for_layout_tool() ↓ (generates production-ready structured frontmatter) 4. create_presentation_from_markdown() ↓ (creates final PowerPoint presentation) ``` This approach transforms LLMs from layout pickers into intelligent presentation consultants.