PPT Master

Overview

PPT Master is an open-source workflow, or skill, for agent-capable AI tools that generates natively editable PowerPoint decks from documents or topics. It produces real PowerPoint objects such as shapes, transitions, animations, data-backed charts and tables, and audio narration from speaker notes rather than flat images. It also supports templates, reuse of existing .pptx files, and multiple generation routes.

Key Features

• Generates natively editable PPTX with real PowerPoint shapes, slide masters, and layouts, not flat images.
• Accepts PDF, DOCX, URL, Markdown, images, and other document formats as source material.
• Supports native transitions, animations, data-backed charts and tables, and audio narration from speaker notes.
• Works with free design, structured templates, and your own .pptx templates.
• Can distill reusable brand, style, layout, and deck templates from references.
• Can fill an existing .pptx with new content while preserving its design and unchanged pages.

Use Cases

• Creating a deck from a PDF or report placed in the projects directory.
• Turning a topic or pasted text into a PowerPoint presentation directly in an AI chat.
• Reusing an existing .pptx template or brand deck for new content.
• Adding native charts, tables, formulas, notes, animations, or narration to a finished deck.

Getting Started

• Install Python 3.10+, clone or download the project, run pip install -r requirements.txt, open the project folder in an agent-capable AI tool, place source files in projects/, and ask the AI to create a deck from a chosen file or pasted content. The project notes that no coding

Deployment & Requirements

• Python 3.10+ and an agent-capable AI tool that can read and write files, execute commands, and sustain multi-turn conversation. Install dependencies with pip install -r requirements.txt. Pandoc is optional and only needed for legacy document formats such as .doc, .odt, .rtf,

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• The project states that the model sets the ceiling, so output quality depends on the chosen model.
• The tool is intended to remove tedious work, but users should not expect a finished, perfect deck in one shot.
• Privacy: apart from AI model communication, the pipeline runs locally on the user's machine.
• Capability boundary: SmartArt is intentionally omitted.