A new open-source project launched on GitHub on June 7, 2026, addresses a longstanding limitation in AI-generated presentations: the inability to edit outputs. GordenSuperPPTSkills gained 543 stars within two days by generating PowerPoint presentations from GPT image output, then converting them into fully editable native PowerPoint files.
Three-Module System Enables End-to-End Automation
The system divides presentation generation into three modular components that work independently or in sequence:
- GordenImagePPTGen: Generates presentation slides as images using GPT's image generation capabilities
- GordenImage2PPTX: Converts image-based slides into editable PowerPoint by extracting four layers: background, structural framework, icons/decorations, and text
- GordenSuperPPTSkill: Orchestrates both skills for complete automation from prompt to editable file
The conversion process leverages GPT's vision capabilities to intelligently parse visual elements, then reassembles them as native PowerPoint objects with coordinates and formatting preserved. Users see intermediate processing images during conversion, providing visual feedback on the transformation.
Layer-Based Reconstruction Produces Native PowerPoint Objects
Unlike traditional AI presentation tools that export static images or PDFs, GordenSuperPPTSkills produces genuinely editable output. The layer-based reconstruction approach separates background elements from structural components, icons, and text, enabling precise reassembly as modifiable PowerPoint objects.
The project's modular architecture allows flexible deployment—users can employ individual skills separately or combine them for full automation. The system is designed specifically for Codex, an AI agent platform, to leverage GPT's multimodal capabilities.
Rapid Adoption Signals Demand for Editable AI Presentations
The project's rapid star growth from 543 stars in 48 hours suggests strong demand for AI presentation tools that produce truly editable output rather than static exports. The creator describes it as the "AI PPT track terminator, the strongest PPT Skill in history," reflecting confidence in the technical approach.
Key Takeaways
- GordenSuperPPTSkills gained 543 GitHub stars within two days of launching on June 7, 2026
- The system converts GPT-generated presentation images into fully editable PowerPoint files using layer-based reconstruction
- Three modular components work independently or together: image generation, image-to-PPTX conversion, and orchestration
- The conversion process extracts four layers—background, structure, icons, and text—then reassembles them as native PowerPoint objects
- Rapid adoption indicates strong demand for AI presentation tools that produce genuinely editable output