<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> This Codex skill converts static slide images, PDFs, and image-based PPTX files into editable PowerPoint decks. It targets users needing to remix existing visual presentations, restoring text, shapes, and images as distinct, adjustable assets.

How It Works

The skill employs a multi-agent, iterative approach. Each input page is processed by a dedicated "page worker" that reconstructs it into editable components: native text boxes, PowerPoint shapes, and separate image assets. This involves cycles of reconstruction, self-checking, and in-page correction for high fidelity. Text size and position are precisely measured via OCR annotations, ensuring consistent styling.

Quick Start & Requirements

• Install: The editppt CLI is automatically installed by the AI during skill execution. Usage involves invoking the skill within a compatible agent (e.g., $image-to-editable-ppt in Codex).
• Prerequisites: Requires an environment supporting agent dispatch and file I/O. A Baidu AI Studio OCR Token is highly recommended for accurate text restoration; otherwise, it falls back to offline detection with reduced quality. Codex OAuth or an OpenAI-compatible API configuration is needed for image backend operations.
• Resource Footprint: High token consumption is a significant factor; restoring a 10-page PPT may consume substantial API quota and take over 10 minutes per page.
• Links: GitHub: https://github.com/ningzimu/image-to-editable-ppt-skill, Baidu AI Studio Token: https://aistudio.baidu.com/account/accessToken.

Highlighted Details

• Supports single/multiple images, multi-page PDFs, and image-based PPTX inputs.
• Pages are processed in parallel by page worker/subagent instances.
• Text elements are precisely sized and positioned based on OCR measurements.
• Preserves original page order and PPTX page remarks.
• Handles complex pages by separating editable text, simple shapes, and image assets.

Maintenance & Community

• Community: A QR code links to a "Skill Exchange Group" for user interaction, feedback, and updates.

Licensing & Compatibility

• License: MIT.
• Compatibility: The MIT license permits commercial use and integration into closed-source projects.

Limitations & Caveats

This is not a lightweight converter, incurring substantial token costs and processing time, making it best suited for ChatGPT Pro users. Effectiveness is not guaranteed for models below GPT-4.5. Minor visual offsets may occur. Complex elements like photos or structured data are often preserved as static image assets rather than fully editable components.