ppt-master

PPT Master is an open-source tool that uses AI to generate fully native, editable PowerPoint (.pptx) presentations from documents, PDFs, DOCX files, web pages, or direct text input. Unlike tools that flatten output into images, PPT Master produces real PowerPoint objects including native shapes with adjustment handles, data-backed charts and tables, slide masters, transitions, animations, and audio narration from speaker notes. The tool runs as a workflow inside any agent-capable AI IDE or chat tool such as Claude Code, GPT, Gemini, Cursor, Codex CLI, or Kimi. The user installs Python and the project dependencies, then instructs the AI agent to generate a deck from source material. All processing occurs locally; only AI model API calls go external. There is no platform subscription or model lock-in. Key capabilities include: - Converting PDFs, DOCX, HTML, EPUB, Jupyter notebooks, and other formats into structured presentations - Custom template support: users can provide their own .pptx templates which are preserved byte-for-byte on unchanged pages - Data-backed charts and tables with editable data via the --native-charts-and-tables flag - Formulas rendered as editable OMML for PowerPoint 2010+ - Optional quick generation mode that skips design confirmation for faster output - Extraction and reuse of brand, style, layout, and deck templates from reference materials - Narration from speaker notes and native animation support The project is MIT-licensed, free to use, and maintained by Hugo He. Recommended models for best results include Kimi K3 or Claude with large context windows, paired with AI image generation tools such as gpt-image-2 or gemini-3.1-flash-image. The README notes that results improve with higher-quality models and that polishing may still be needed after generation.