PPT Master turns documents into native PowerPoint
The MIT-licensed open-source project builds real PPTX files from documents, with native shapes, animations, charts and speaker-note audio.
An empty conference room at dusk with a faintly glowing projection screen — scene generated by an AI image model.
PPT Master is an open-source Python tool that converts a document or a bare topic into an editable PowerPoint file made of native shapes, transitions, charts and tables rather than flattened slide images.
At a glance
- MIT license, written in Python, minimum version listed on the project page is Python 3.10.
- Project page on August 31, 2026: about 50,400 stars, about 4,100 forks, 6 open issues, 1,852 commits on main.
- SVG serves as the intermediate format; output is translated to DrawingML, PowerPoint's native drawing format.
- Recommended models are Kimi K3 or Claude with roughly 1 million tokens of context, plus gpt-image-2 or gemini-3.1-flash-image.
- Speaker notes can be turned into an audio track, and custom .pptx templates are supported.
PPT Master is an open-source Python tool that turns a document, or just a topic, into an editable PowerPoint file. What separates it from many deck generators is the output format: the project says it emits native shapes, connectors and adjustment handles instead of flat rendered images. Anyone who opens the resulting file can still edit text, fills and effects.
What it produces
The project page lists slide transitions and animations across four categories: entrance, emphasis, motion path and exit. Charts and tables can be generated as native PowerPoint objects rather than embedded pictures. Speaker notes can be converted into an audio narration track. Custom .pptx templates are supported, so output can follow an existing house style.
Under the hood the pipeline uses SVG as an intermediate format, then translates to DrawingML, the drawing format PowerPoint uses internally. The project describes itself as running locally, with no platform lock-in.
Requirements and workflow
According to the project page, Python 3.10 or newer is enough, along with the dependencies in requirements.txt. The repository is then opened inside an agent-capable AI tool; Claude Code, Cursor and GitHub Copilot are named. Users drop in source material such as PDF, DOCX or images, or paste content directly. The project summarizes the workflow as three steps: install Python, install an AI tool, add your material.
For model choice, the project recommends Kimi K3 or Claude with a large context window of roughly 1 million tokens, paired with an image model such as gpt-image-2 or gemini-3.1-flash-image. Other models reportedly work, with trade-offs in output quality.
How widely it is used
When the page was retrieved on August 31, 2026, it showed about 50,400 stars and about 4,100 forks, alongside 6 open issues and 1,852 commits on the main branch. Those counters are a snapshot and move daily. The license is MIT.
What we could not verify
Exactly one independent source was available for this article: the project page itself. Every claim about features, workflow and model recommendations therefore comes from the maintainer and has not been checked against a second newsroom. In particular, we cannot confirm how reliably the generated files open across PowerPoint versions, or how good the results are in day-to-day use.
The project itself tempers expectations, calling the software a tool whose output depends on the model behind it and still needs polish. We have no adoption figures beyond the counters displayed on the project page.
FAQ
Is PPT Master free to use?
The source code is published under the MIT license, so the software itself is free. Costs come from the AI model and image model you connect yourself.
Which Python version does PPT Master need?
The project page lists Python 3.10 as the minimum, plus the dependencies in requirements.txt.
Can the generated slides be edited afterwards?
The project states that it creates native PowerPoint objects, meaning shapes, charts and tables rather than fixed images, so text, fills and effects stay editable. We have not verified this independently.
