⚠️ Status: early development. This library is not yet functional. This README describes the intended scope so early visitors know what's coming. Check back for releases.
A standalone Python library for extracting structured content from PowerPoint (.pptx) files with a focus on embedded charts and SmartArt, two content types most extraction tools miss entirely. No LibreOffice or other external binary required. Built for feeding PPTX content into RAG pipelines and document processing systems and designed to handle large decks (500+ slides) without memory failures.
Extracting structured data from PowerPoint files is a common need for RAG pipelines and document processing systems, but existing tools have gaps, especially when it comes to embedded charts and SmartArt. We tested a number of libraries and found that while some handle parts of this processing well, none were able to fully extract the content.
There are also memory and performance issues when dealing with large decks (500+ slides). Additionally, some libraries require LibreOffice to work correctly, which isn't always permissible in production environments.
- Standard text and table extraction
- Embedded chart extraction: including both classic and modern (ChartEx) chart formats, with data and headers intact.
- SmartArt extraction: pulling text content out of SmartArt nodes, which no tested alternative currently supports.
- Large deck handling: designed to process decks of 500+ slides without memory failure.
- No external dependencies — A fully-functional open-source library without LibreOffice dependancy.
Not yet published. Once an initial version is available, install instructions (likely via PyPI) will go here.
Not open to contributions yet while the core shape of the library is being established. Watch this repo for updates — issues and discussion will open once there's a working baseline.
Want to know when this ships? Email signup to follow.
TBD, will be finalized before first release.