Integrated Microsoft Office Control Modules (MCP)

This repository furnishes dedicated MCP servers facilitating AI-driven interaction with core Microsoft Office applications, namely PowerPoint and Excel. These tools enable sophisticated content generation, data manipulation, and structural analysis using natural language directives.

Both server modules rely on the pywin32 library for direct Component Object Model (COM) automation, ensuring deep integration with active, running Office processes.

System Requirements

• OS: Windows required for COM interface access.
• Software: Installed copies of Microsoft Office (PowerPoint and/or Excel).
• Runtime: Python version 3.7 or later.
• Package: The pywin32 library must be installed.

Setup Procedure

1. Acquire Codebase: bash git clone https://github.com/jenstangen1/mcp-pptx.git cd mcp-pptx

2. Dependency Installation (Recommended: using uv): bash uv pip install pywin32

3. COM Initialization: Execute the pywin32 post-installation script with elevated permissions: bash python C:\path\to\your\env\Scripts\pywin32_postinstall.py -install

Configuration for Claude Integration

Integrate these services into your Claude Desktop client by updating your configuration JSON with the following structure, adjusting the cwd to your repository root:

{ "mcpServers": { "powerpoint_mcp_win32": { "command": "uv", "args": [ "run", "mcp_powerpoint_server_win32.py" ], "cwd": "C:\path\to\your\workspace" }, "excel_mcp_win32": { "command": "uv", "args": [ "run", "mcp_excel_server_win32.py" ], "cwd": "C:\path\to\your\workspace" } } }

---

PowerPoint Interaction Module

This module exposes an extensive API enabling large language models to programmatically construct, refine, and populate PowerPoint deck content, with specialized capabilities for visualizing financial datasets.

Core Capabilities

Presentation Lifecycle Management

• Creation, modification, and persistence of .pptx files.
• Slide addition, removal, and structural modification.
• Workspace-relative file loading and saving mechanisms.
• Support for saving standardized presentation layouts as reusable templates.

Graphical Object Manipulation

• Precise control over text boxes, shapes, embedded images, and embedded charts.
• Advanced geometric arrangement, grouping, and layering of visual components.
• Automated rendering of connecting lines between specified geometric elements.

Financial Data Visualization

• Generation of diverse chart types (bar, pie, waterfall, line) populated with financial indicators.
• Data handling for core metrics: Revenue, Profit, EBITDA, Assets, Equity, Margins, and Growth Rates.
• Ability to structure key metrics into tabular formats for cross-entity comparison.
• Note: Currently employs synthetic datasets; future integration targets the Proff API for live Norwegian corporate data.

Aesthetic Control

• Application of complex styling attributes to text runs and shapes (e.g., gradients, borders, fills).
• Global presentation styling adjustments (e.g., slide background properties).

Available Service Calls (Tools)

Category	Tool Name	Function Description
File Ops	list_presentations	Enumerates presentation files in the defined work area.
	upload_presentation	Ingests a local file into the managed workspace.
	save_presentation	Commits current presentation state to disk.
Slide Control	add_slide	Inserts a new, blank slide.
	delete_slide	Removes a specified slide index.
	get_slide_count	Reports the total slide count.
	analyze_slide	Extracts structured content data from a slide.
	set_background_color	Defines the canvas color for the active slide.
Element Control	add_text	Places and formats text onto a slide.
	add_shape	Draws a basic geometric object.
	edit_element	Modifies the positional or inherent properties of an existing object.
	style_element	Applies specific visual appearance settings to an object.
	connect_shapes	Draws a line object linking two existing shapes.
	find_element	Locates objects based on attributes (e.g., text content, type).
Financial Ops	get_company_financials	Fetches structured financial data (currently mocked).
	create_financial_chart	Renders a chart object based on provided data/type.
	create_comparison_table	Generates a grid summarizing metrics across entities.
Template Mgmt	list_templates	Lists available saved layout definitions.
	apply_template	Applies a layout definition to the current presentation.
	create_slide_from_template	Generates a new slide based on a specific template asset.
	save_as_template	Archives the current slide structure as a reusable template.
Debugging	debug_element_mappings	Provides introspection into how elements map to COM identifiers.

---

Excel Automation Module

This module equips AI agents with the capability to interact deeply with open or newly created Excel workbooks, focusing on data extraction, computational tasks, and structural organization.

Feature Set

Workbook Administration

• Establishing connections to live Excel application processes.
• Retrieving a roster of all actively opened document files.
• Saving workbooks with explicit control over file format extensions (.xlsx, .xlsm, etc.).

Worksheet Structure

• Listing all sheets within the active workbook.
• Procedures for instantiating new workbook sheets.
• Direct referencing of worksheets via their naming convention or sequential position.

Data Cell Interaction

• Reading the content of a single specified cell.
• Writing a new value to an individual cell address.
• Bulk reading of data spanning defined cell regions (ranges).
• Bulk writing of data arrays into specified cell regions.
• Automatic handling and casting for numerical, date, and currency data types.

Available Service Calls (Tools)

Category	Tool Name	Function Description
Workbook	list_open_workbooks	Displays all Excel files currently in memory.
	save_workbook	Persists the active workbook state to disk, allowing format selection.
Worksheet	list_worksheets	Returns names of all sheets in the active workbook.
	add_worksheet	Creates and names a new sheet tab.
	get_worksheet	Retrieves a worksheet object reference by name or index.
Cell/Range	get_cell_value	Retrieves the primitive value from one cell.
	set_cell_value	Assigns a new value to one cell.
	get_range_values	Extracts a 2D array of data from a specified range.
	set_range_values	Overwrites a region with a 2D array of input data.

---

Operational Notes

• Platform Lock-in: Both servers are strictly dependent on the Windows OS environment and active Office installation.
• Execution Model: The tools operate by interfacing with live, instantiated Office processes.
• Troubleshooting COM: If errors related to the Component Object Model arise, re-running the pywin32_postinstall.py script is the primary remedy.

---

Usage Guidance: Interacting with Claude

Once configuration is complete, natural language prompts translate directly into sequential tool executions.

Example Flow (Presentation): Input: "Design a presentation on market penetration showing Q3 vs Q4 revenue in a bar chart." Execution Chain: add_slide -> add_text (Title) -> create_financial_chart (Revenue Bar Chart).

Example Flow (Spreadsheet): Input: "In 'SalesData.xlsx', find the highest value in Column C of the 'Totals' sheet and report it in cell D1 of a new sheet named 'MaxFinder'." Execution Chain: get_worksheet (Totals) -> get_range_values (Column C) -> set_range_values (New Sheet D1, incorporating calculated max value).