image2ppt

Convert slide images and scanned PDFs into object-level editable PowerPoint files.

39|2|Updated Aug 16, 2026
One-click install
npx skills add https://github.com/Paul-Jeo/Image2PPT --skill image2ppt-paul-jeo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: image2ppt
Source: https://github.com/Paul-Jeo/Image2PPT
Command: npx skills add https://github.com/Paul-Jeo/Image2PPT --skill image2ppt-paul-jeo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pypdfium2, Pillow, numpy, requests, PyYAML, openai, and includes scripts (resource) and references (resource) components.

What problem does it solve? Rebuilding a slide from a screenshot or scanned PDF normally means retyping text and redrawing every shape by hand. This Skill reconstructs complex business slides and scientific flowcharts as native, individually editable PowerPoint objects instead of pasting a flat screenshot. ## Core Features & Use Cases - Object-level reconstruction: Text, shapes, circle nodes, connectors, and arrows are rebuilt as independent native PowerPoint objects with measured positions, not as a background image. - Compound diagram handling: Knowledge graphs, flowcharts, and dense timelines are decomposed into semantic regions with measured nodes, relations, and single-object arrows. - Rendered QA gates: Each page and the final deck pass visual-review evidence checks, arrow atomicity inspection, and speaker-note preservation before delivery. - Use Case: You have a scanned PDF of a research framework diagram. The Skill measures every node and connector, rebuilds them as native shapes, renders the result, and delivers an editable PPTX with a QA report. ## Quick Start Use the image2ppt skill to convert these slide images into an editable PPTX while keeping all text, nodes, and arrows as native PowerPoint objects.

Frequently Asked Questions about image2ppt

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert a slide image into an editable PowerPoint?▼

Provide one or more ordered PNG/JPG files, a scanned PDF, or an image-only PPTX, then run the image2ppt prepare command to create a run directory. Each page is reconstructed as native text boxes, shapes, and connectors, validated, and finalized into an editable PPTX.

Can image-to-PPT conversion keep flowchart nodes and arrows editable?▼

Yes. Compound diagrams are measured node by node: circles become native ellipse shapes, relations become native line connectors, and arrows are single connector or AutoShape objects with arrowheads, never a line plus a separate triangle.

Does image2ppt work without a PaddleOCR token?▼

Yes, but with reduced capability. Without a Baidu AI Studio PaddleOCR token it falls back to the builtin-ink offline detector, which measures text region geometry but does not recognize the actual characters.

What input formats does image2ppt support?▼

It accepts one or more ordered PNG or JPG images, scanned PDFs, and image-only PPT or PPTX files. Inputs are normalized into per-page source images inside a run directory before reconstruction begins.

What are the limitations of image-based PPT reconstruction?▼

It reconstructs existing visual pages and does not author new decks from notes or outlines. Complex illustrations or photos that cannot be measured reliably remain as bounded replaceable image assets, and low-resolution sources or missing fonts reduce fidelity.