This project addresses the challenge of rapidly generating presentation-ready slides from diverse source materials like research papers, code repositories, and documents. It targets engineers, researchers, and power users by automating the creation of natively editable PowerPoint (.pptx) files, ensuring accuracy through source tracing and incorporating an independent critic review, thereby streamlining the presentation development workflow.

How It Works

Slide-maker employs a multi-agent architecture, comprising specialized roles like a Content-planner, Slide-design (art director), and an independent Critic. This approach ensures that content is meticulously read from source materials (papers, repos, docs, etc.) without fabrication. The design is then composed slide-by-slide, focusing on content integration rather than mere template filling. A crucial differentiator is the critic agent, which reviews generated slides for content fidelity and design flaws, providing an objective quality gate before delivery. The output is a fully editable native PowerPoint file, not flattened images.

Quick Start & Requirements

• No Install Option: Utilize the slide-maker GPT directly within ChatGPT via the GPT Store for zero setup.
• Local Install:
  • Clone the repository: git clone --depth 1 https://github.com/addsumtech/slides_maker.git
  • Install Python dependencies: cd slides_maker && python3 -m pip install -r skills/slide-maker/requirements.txt
  • Register the skill: python3 skills/slide-maker/scripts/install_skill.py --target both (or codex/claude). Alternatively, use npx skills add addsumtech/slides_maker.
• Prerequisites: Python 3.9+, LibreOffice (for previews), and an SVG rasterizer (librsvg, cairosvg, or Chrome/Edge headless). OS-specific installation instructions are provided.
• Invocation: Interact via a step-by-step interview (/slide-maker command or plain text) or use one-liner commands.
• Links: slide-maker GPT, Troubleshooting & FAQ.

Highlighted Details

• Native Editable PPTX: Generates real text boxes, shapes, native charts, and equations that are fully editable within PowerPoint/Keynote.
• Source-Traced Content: Guarantees that all data, numbers, and figures originate from the provided source material, preventing AI-generated inaccuracies.
• Independent Critic Review: An integrated critic agent validates content accuracy and design quality, ensuring a polished output.
• Multi-Format Output: Supports recomposition for various aspect ratios including 16:9, 4:3, 1:1, 3:4 (Redbook), 9:16 (story covers), and A4 print, each with appropriate layout logic.
• Iterative Refinement: Generated decks can be cleanly rebuilt and tuned via chat commands based on user feedback.
• Input Versatility: Accepts papers, code repos, docs, Word, Excel, CSV, images, video/audio, books, or just a topic (which it researches online).

Maintenance & Community

The project is developed by Addsum. Specific details regarding active maintenance, community channels (like Discord/Slack), or notable contributors are not detailed in the provided README.

Licensing & Compatibility

The project is released under the MIT license. This permissive license allows for commercial use and integration within closed-source projects without significant restrictions.

Limitations & Caveats

The tool does not offer zero-setup cloud deployment or shareable links; it produces a local file. It does not generate animated web backgrounds. While it aims for high quality, it does not guarantee a perfect deck in a single pass, emphasizing that the final polish is the user's responsibility. AI usage costs apply, scaling with the complexity and thoroughness of the review process.