PPTX MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@PPTX MCP Server create a 3-slide presentation about quarterly sales results"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
š PPTX MCP Server
A Model Context Protocol (MCP) server that enables AI assistants to create, edit, and manipulate PowerPoint presentations programmatically.
⨠Features
Full PowerPoint Control - Create, read, and modify
.pptxfiles without needing PowerPoint installedAI-Native Design - Built specifically for LLMs to generate and edit presentations through structured JSON
Rich Formatting Support - Text styling, colors, alignment, bullets, shapes, and backgrounds
Template Workflows - Extract content from existing presentations, modify, and regenerate
Visual Debugging - Generate thumbnail grids to preview slides programmatically
Office XML Access - Direct access to underlying OOXML for advanced customization
Related MCP server: mcp-powerpoint
š Quick Start
1. Install
# Clone the repository
git clone https://github.com/YOUR_USERNAME/pptx-mcp-server.git
cd pptx-mcp-server
# Install the package
pip install -e .2. Configure Your MCP Client
Add to your MCP settings:
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"pptx": {
"command": "python",
"args": ["-m", "pptx_mcp_server"]
}
}
}Add to your MCP configuration:
{
"mcpServers": {
"pptx": {
"command": "python",
"args": ["-m", "pptx_mcp_server"]
}
}
}Specify the full path to your Python interpreter:
{
"mcpServers": {
"pptx": {
"command": "/path/to/your/python",
"args": ["-m", "pptx_mcp_server"]
}
}
}Examples:
Conda:
/Users/username/miniconda3/bin/pythonvenv:
/path/to/project/.venv/bin/python
3. Restart Your MCP Client
Restart Claude Desktop, Cursor, or your MCP client to load the server.
š¦ Requirements
Python 3.10+
Dependencies (installed automatically):
mcp- Model Context Protocol SDKpython-pptx- PowerPoint file manipulationPillow- Image processinglxml- XML parsingdefusedxml- Secure XML parsing
Optional (for thumbnails)
# macOS
brew install --cask libreoffice
brew install poppler
# Ubuntu/Debian
sudo apt-get install libreoffice poppler-utilsš ļø Available Tools
Tool | Description |
| Create new presentations from scratch |
| Extract text content with positions and formatting |
| Replace text using JSON specifications |
| Duplicate, delete, and reorder slides |
| Generate visual thumbnail grids |
| Extract Office files to editable XML |
| Rebuild Office files from XML |
| Validate document structure |
š Usage Examples
Create a New Presentation
{
"output_path": "/path/to/presentation.pptx",
"layout": "16:9",
"slides": [
{
"background": "#0f172a",
"shapes": [
{
"type": "textbox",
"left": 0.5,
"top": 3,
"width": 12,
"height": 1.5,
"text": "Welcome to My Presentation",
"font_size": 54,
"bold": true,
"color": "#ffffff",
"alignment": "center"
},
{
"type": "textbox",
"left": 0.5,
"top": 5,
"width": 12,
"height": 1,
"text": "Subtitle goes here",
"font_size": 24,
"color": "#94a3b8",
"alignment": "center"
}
]
},
{
"shapes": [
{
"type": "textbox",
"left": 0.5,
"top": 0.5,
"width": 12,
"height": 1,
"text": "Key Points",
"font_size": 36,
"bold": true
},
{
"type": "textbox",
"left": 0.5,
"top": 1.8,
"width": 12,
"height": 5,
"paragraphs": [
{"text": "First important point", "font_size": 24, "bullet": true},
{"text": "Second important point", "font_size": 24, "bullet": true},
{"text": "Third important point", "font_size": 24, "bullet": true}
]
}
]
}
]
}Supported shape types:
textbox- Text contentrectangle- Rectangle (can contain text)rounded_rectangle- Rounded cornersoval- Circle/ellipseimage- Image file (usepathproperty)line- Line connector
Supported layouts: 16:9, 4:3, widescreen, standard
Extract Text Inventory
Get all text content from an existing presentation:
{
"pptx_path": "/path/to/presentation.pptx"
}Returns structured JSON:
{
"slide-0": {
"shape-0": {
"left": 0.5,
"top": 1.0,
"width": 12.0,
"height": 1.5,
"paragraphs": [
{"text": "Title Text", "font_size": 44.0, "bold": true}
]
}
}
}Replace Text Content
Modify text in an existing presentation:
{
"pptx_path": "/path/to/template.pptx",
"output_path": "/path/to/output.pptx",
"replacements_json": {
"slide-0": {
"shape-0": [
{"text": "New Title", "font_size": 44, "bold": true}
]
}
}
}Rearrange Slides
Reorder, duplicate, or remove slides:
{
"template_path": "/path/to/template.pptx",
"output_path": "/path/to/output.pptx",
"slide_sequence": "0,2,2,1,3"
}0,2,2,1,3ā Keep slide 0, duplicate slide 2, then slides 1 and 3Omit an index to delete that slide
Unpack/Pack for XML Editing
// Unpack to directory
{
"office_file": "/path/to/document.pptx",
"output_dir": "/path/to/unpacked"
}
// Pack back to file
{
"input_dir": "/path/to/unpacked",
"output_file": "/path/to/output.pptx"
}š§ Troubleshooting
Ensure you're using the correct Python environment:
# Check which Python pip uses
pip --version
# Install with specific Python
/path/to/python -m pip install -e .Verify the config file path is correct for your OS
Ensure JSON syntax is valid (no trailing commas)
Restart the MCP client completely
Check logs for errors
Install LibreOffice and poppler:
# macOS
brew install --cask libreoffice && brew install poppler
# Linux
sudo apt-get install libreoffice poppler-utilsEnsure the output paths are writable and parent directories exist.
š Project Structure
pptx-mcp-server/
āāā pyproject.toml # Package configuration
āāā README.md
āāā pptx_mcp_server/
āāā __init__.py
āāā server.py # MCP server implementation
āāā tools/
āāā __init__.py
āāā create.py # Create new presentations
āāā inventory.py # Extract text content
āāā replace.py # Text replacement
āāā rearrange.py # Slide manipulation
āāā thumbnail.py # Visual thumbnails
āāā ooxml.py # XML pack/unpack/validateš¤ Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Fork the repository
Create your feature branch (
git checkout -b feature/AmazingFeature)Commit your changes (
git commit -m 'Add some AmazingFeature')Push to the branch (
git push origin feature/AmazingFeature)Open a Pull Request
š Acknowledgments
Built with python-pptx
Uses the Model Context Protocol by Anthropic
Available Tools
8 toolsapply_text_replacementsA
Apply text replacements to a PowerPoint presentation using a JSON specification. The JSON should map slide/shape IDs to new paragraph content with formatting. All text shapes are cleared unless explicitly provided with new content.
Inputs: pptx_path (required), replacements_json (required), output_path (required).
create_presentationA
Create a new PowerPoint presentation from scratch. Accepts a JSON specification with slides, shapes, text content, and formatting. Supports text boxes, rectangles, ovals, images, and various text formatting options. Use this to create new presentations before using other tools to modify them.
Inputs: output_path (required), layout, slides.
create_thumbnail_gridA
Create visual thumbnail grids from PowerPoint slides. Useful for quick visual analysis of presentation structure and layouts. For large presentations, multiple grid images are created automatically.
Inputs: pptx_path (required), output_prefix, cols, outline_placeholders.
extract_text_inventoryA
Extract structured text content from a PowerPoint presentation. Returns JSON with all text shapes, their positions, and formatting details. Useful for understanding presentation structure before making replacements.
Inputs: pptx_path (required), output_path, issues_only.
pack_office_documentA
Pack a directory back into an Office document (.docx, .pptx, .xlsx). Removes pretty-printing whitespace from XML before packing. Can optionally validate the document before saving.
Inputs: input_dir (required), output_file (required), validate, force.
rearrange_slidesB
Rearrange slides in a PowerPoint presentation. Can duplicate, delete, and reorder slides based on a sequence of indices. Slide indices are 0-based. The same index can appear multiple times to duplicate.
Inputs: template_path (required), output_path (required), slide_sequence (required).
unpack_office_documentA
Unpack an Office document (.docx, .pptx, .xlsx) to a directory. The XML files are pretty-printed for easy reading and editing. Use this to inspect or manually edit the raw XML structure.
Inputs: office_file (required), output_dir (required).
validate_office_documentA
Validate an unpacked Office document against XSD schemas. Checks XML well-formedness, namespace declarations, unique IDs, file references, content types, and schema compliance. Returns detailed error messages for any issues found.
Inputs: unpacked_dir (required), original_file (required), verbose.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelogā¦
TDQS
Scored across 8 tools
Most tools have distinct purposes focused on PowerPoint operations, but there is some overlap between pack_office_document and unpack_office_document with validate_office_document, as all three handle Office document file structures. However, their specific functions (packing, unpacking, validating) are clearly differentiated in descriptions, minimizing confusion.
All tool names follow a consistent verb_noun pattern using snake_case, such as apply_text_replacements, create_presentation, and extract_text_inventory. This uniformity makes the tool set predictable and easy to navigate for agents.
With 8 tools, the server is well-scoped for PowerPoint manipulation, covering creation, modification, extraction, and file handling. Each tool serves a clear purpose without redundancy, fitting the domain effectively.
The tool set provides comprehensive coverage for PowerPoint operations, including creation, text editing, slide management, and file handling. A minor gap exists in advanced formatting or animation tools, but core workflows are fully supported, allowing agents to perform most common tasks.
Related MCP Connectors
Generate AI slide decks and export native PowerPoint (.pptx) and PDF presentations from chat.
Create, edit, export and manage presentations. Requires a paid PPT.AI API merchant account.
Generate, edit, merge, translate and PDF-convert PowerPoint (.pptx) over MCP. 8 tools.
Create professionally designed, editable PowerPoint presentations from a prompt.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to programmatically create, manipulate, and analyze Microsoft PowerPoint presentations with advanced formatting and template management.5MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to create and manipulate PowerPoint presentations programmatically, including adding slides, exporting to PDF, and reading metadata.16 npm5MIT
- FlicenseNot gradedqualityDmaintenanceEnables AI assistants to create, modify, and manage PowerPoint presentations programmatically using python-pptx.17-
- AlicenseNot gradedqualityBmaintenanceEnables AI assistants to automate Microsoft PowerPoint through natural language, supporting slide creation, text formatting, charts, animations, and export via the COM API.42MIT