powerpoint-automation-mcp

# MCP Server Demo - PowerPoint Automation This project demonstrates how to use MCP (Model Completion Protocol) to create a server that provides PowerPoint automation capabilities for Amazon Q. ## Demo https://github.com/user-attachments/assets/3daf3bef-4d75-4639-a891-0e64b80b4807 ## Features - Create, open, and save PowerPoint presentations - Add and modify slides - Add text boxes, images, tables, and charts - Update text content and shape properties - Retrieve table content ## Running the Server To run the server: ``` uv run --directory /path/to/mcp-ppt-server mcp-ppt-server ``` ## Usage The server exposes various tools for PowerPoint automation that can be called via MCP. The server is designed to work with a single active presentation at a time, which simplifies the API by eliminating the need to track and specify presentation IDs. ### Presentation Management - `initialize_powerpoint()`: Initialize PowerPoint automation - `create_presentation(template=None)`: Create a new presentation, optionally using a template file - `open_presentation(file_path)`: Open an existing presentation from the specified path - `get_presentation()`: Get information about the currently active presentation - `save_presentation(path=None)`: Save the active presentation to disk (path is optional if presentation was previously saved) - `close_presentation()`: Close the active presentation ### Slide Management - `get_slides()`: Get a list of all slides in the active presentation - `add_slide(layout_index=1)`: Add a new slide with specified layout (default is Title and Content) - Layout options: 0=Title slide, 1=Title and content, 2=Section header, 3=Two content, etc. - `set_slide_title(slide_index, title)`: Set the title text of a slide ### Content Management - `get_slide_text(slide_index)`: Get all text content in a slide - `get_slide_shapes(slide_index)`: Get all shapes in a slide with their IDs and properties - `update_text(slide_index, shape_index, text)`: Update the text content of a shape - `update_shape_by_id(slide_index, shape_id, text=None, left=None, top=None, width=None, height=None)`: Update a shape by its ID with new properties - `update_table_cell(slide_index, shape_index, row, col, text)`: Update the text in a table cell ### Adding Content - `add_textbox(slide_index, text, left=1, top=1, width=4, height=2)`: Add a text box to a slide - `add_image(slide_index, image_path, left=1, top=1, width=None, height=None)`: Add an image to a slide - `add_table(slide_index, rows, cols, left=1, top=1, width=8, height=4)`: Add a table to a slide - `get_table_content(slide_index, shape_index)`: Retrieve the content of a table in a slide - `add_chart(slide_index, chart_type, categories, series_names, series_values, left=1, top=1, width=8, height=4, has_legend=True)`: Add a chart to a slide with multiple data series - Chart types: 'COLUMN', 'LINE', 'PIE', 'BAR' - `categories`: List of category names (x-axis labels) - `series_names`: List of series names for the legend - `series_values`: List of lists containing values for each series ## Error Handling Most functions return a dictionary with either success information or error details. Always check for the presence of an "error" key in the returned dictionary to handle errors appropriately. Example error response: ```json { "error": "No active presentation. Please open or create a presentation first." } ``` ## MCP Config ```json { "mcpServers": { "ppt-mcp-server": { "command": "uv", "args": [ "run", "--directory", "/path/to/mcp-ppt-server", "mcp-ppt-server" ] } } } ```