content / skill

PPT Master

Ask your coding agent to turn documents or a topic into an editable PowerPoint deck.

At a glance

Purpose
Ask your coding agent to turn documents or a topic into an editable PowerPoint deck.
Runs in
Local agent with shell
Operating systems
macOS, Linux, Windows
Needs
Runtime: Python 3.10+; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data
Reads files: Yes; Writes files: Yes; Sends data to other services: Yes
Tool type
Skill
Delivery
From GitHub
Version
Latest on the main branch
Access
Code on GitHub
Compatibility
According to the README, it runs in any tool with agent capability. It names IDE agents such as Cursor, Trae, Codebuddy IDE, Windsurf and Zed, IDE extensions such as Claude Code, GitHub Copilot and Cline, and CLI agents such as Claude Code, Codex CLI and Gemini CLI. It names Claude Code as the author's pick. The README says Claude, GPT, Gemini, Kimi and other models can drive it, and it has a separate Windows installation guide.
External costs
The repository is MIT-licensed. According to the README, the project is free and open source, its cost comes from your AI model usage, and it adds no PPT subscription on top. It says image generation uses your agent's own image tool or a provider API key, and that web image search works without keys, with free Pexels and Pixabay keys as an option. The README also names sponsor services that sell pay-as-you-go model access.
Permissions
The skills act through your coding agent with the access you give it. According to the README, the agent converts the source files you place in the project folder, writes exports and page previews to folders in the project, and can edit an existing .pptx you give it. The README describes optional scripts that search the web for images or call an image-generation provider with your API key, an update script that pulls the latest version for clone installs, and scripts that read settings and keys from environment variables or from the first .env file found in the working directory, the skill folder, the repository root or the .ppt-master folder in your home directory.
Usage terms
MIT License, as stated in the repository. See its LICENSE file for the exact terms. The README says the optional PDF converter depends on PyMuPDF, which is licensed under AGPL-3.0, and is needed only for PDF sources.

Code on GitHub

hugohe3/ppt-master

MIT

Python61,137 stars4,838 forksLast pushed 11 Oct 2026 UTC

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

Topics: ai-agent aippt office powerpoint powerpoint-generation ppt pptx presentation

Public GitHub data fetched · 0 days old

View on GitHub ↗

What it does

PPT Master is a workflow, packaged as a skill, that runs inside an AI tool with agent features and exports PowerPoint files. According to the README, you give it a topic or source material such as a PDF, a DOCX file or pasted text, and it produces a natively editable .pptx file. Charts and tables come as editable shapes by default, and an option produces data-backed PowerPoint charts and tables. The GitHub description also mentions transitions, animations and audio narration from speaker notes.

The README says the main route builds a new deck from source documents. Other routes distill reusable brand, style and layout templates from your references, fill an existing .pptx with new content while keeping its design, and add transitions, animations and narration to a finished deck. By default the agent first confirms a design spec with you, and you can ask for quick generation to skip that step.

The README says the pipeline runs on your machine apart from your calls to an AI model. It also says the model sets the quality ceiling, that you should not expect a perfect deck in one shot, and that the remaining polishing is yours.

Who it is for

People who make presentations and want an AI agent to build an editable PowerPoint deck from documents or a topic.

Example tasks

Create a deck from a PDF report in your projects folder.
Turn pasted text into a presentation.
Quickly generate a 5-page deck without a confirmation step.
Fill an existing .pptx with new content and keep its design.
Add notes, transitions, animations or narration to a finished deck.

Getting started

The README gives three routes: clone the repository with Git, download the ZIP, or install through a skill marketplace with npx skills add hugohe3/ppt-master or, in Claude Code, /plugin marketplace add hugohe3/ppt-master and /plugin install ppt-master@ppt-master. In each route you then run pip install -r requirements.txt. Open the project folder in your agent, put source files in the projects folder or paste text, and ask for a deck in chat. For updates on a clone, it gives python3 skills/ppt-master/scripts/update_repo.py. The README has the exact commands.

Limitations

According to the README, it needs Python 3.10 or later, a pip install of the project's requirements file, and an AI tool with agent features that can read and write files, run commands and hold a multi-turn chat. The README says the model sets the ceiling. The README recommends Kimi K3, from a project sponsor, or Claude to drive the pipeline, with an AI image generator. It says not to expect a finished deck in one shot. Pandoc is needed only for legacy document formats. The optional PDF converter depends on PyMuPDF, which the README says is licensed under AGPL-3.0 rather than MIT. The README calls SmartArt a deliberate omission. The repository can change at any time; check its README before you install.

Data practices

According to the README, your data stays local: apart from communication with the AI model, the entire pipeline runs on your machine. The README describes optional web image search and AI image generation, which call outside services when you use them.

Community experience

Not rated yet

Sign in to share your experience.

Comments 0

No comments yet.

Sign in to comment

Report a listing concern