spire-presentation

The Spire.Presentation MCP Server enables AI agents to interact with PowerPoint (Spire.Presentation) files through the Model Context Protocol (MCP) without requiring Microsoft Office installed. It exposes a set of presentation-manipulation tools—such as adding or deleting slides, creating and editing tables, applying styles, saving shapes as images, and creating charts—via an MCP-compatible interface over Server-Sent Events (SSE). This allows automated agents to read, convert (to PDF/HTML/images), and modify slides, shapes, and charts in bulk, in a scalable way.

To use it, start the server with the uv runner as described in the installation section. The server serves an API over HTTP at a configurable port (default 8000). Tools are organized into categories (Basic Operations, Data Processing, Formatting, Advanced Features, and Conversion). An integration example shows how to connect a client or AI assistant (via Cursor IDE) to the server by pointing to the SSE endpoint and supplying a directory for PPT files. Additionally, the server supports SSE-based transport and can be integrated with gateways to route the stream to different clients if needed.

Prerequisites:

• Python 3.10 or higher
• Git

Install and run locally:

1. Clone the repository

git clone https://github.com/eiceblue/spire-presentation-mcp-server
cd spire-presentation-mcp-server

2. Install the server package using uv (Python package runner)

uv pip install -e .

3. Run the MCP server (default port 8000). The service is exposed via Server-Sent Events (SSE).

uv run spire-ppt-mcp-server

Optional: specify a custom port and PPT files directory via environment variables as shown below.

• The server uses SSE for transport; it can be exposed remotely behind gateways or proxies. Consider using a gateway that supports SSE if you plan to access the server from remote clients.
• Environment variables:
  • FASTMCP_PORT: sets the server port (default 8000).
  • PPT_FILES_PATH: directory where PowerPoint files are stored and accessed by the server (default ./ppt_files).
• If you encounter port conflicts, set FASTMCP_PORT to a different value before starting the server.
• The server is MIT licensed and is suitable for development and integration in AI tooling workflows.