Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

image2pptx

Convert images, screenshots, and scientific figures into editable PowerPoint slides.

Semantic SVG reconstruction first. Native PPTX shapes second. Not a screenshot dropped onto a slide.

中文文档

GitHub Repo stars Windows PowerPoint

Demo

Demo 1: Dynamic graph figure

image2pptx before and after comparison

The comparison above shows the source image on the left and a PowerPoint preview exported from the converted deck on the right. The generated deck contained 212 editable shapes, 68 text runs, and 0 embedded pictures.

Demo 2: Architecture figure

image2pptx architecture before and after comparison

The second demo shows a denser architecture figure reconstructed into a PPT-style preview. The preview image is committed for lightweight browsing; generated .pptx files stay local.

To keep the repository fast to browse, demo .pptx files are generated locally and are not committed.

Why image2pptx

Most image-to-PowerPoint workflows stop at one of two bad outcomes:

  • the original image gets embedded as a flat picture
  • the converted slide looks acceptable but is painful to edit

image2pptx takes a different route. It rebuilds diagrams as semantic SVG first, then converts that SVG into PowerPoint-native text, shapes, arrows, and groups. The result is a .pptx that stays much more editable for manual cleanup and restyling.

Features

  • Convert screenshots, diagram images, scientific figures, and PDF-rendered pages into editable slides
  • Reconstruct text, boxes, arrows, legends, connectors, and icons as native PowerPoint objects
  • Keep a semantic SVG as an intermediate source of truth
  • Export preview images for SVG and PPT output QA
  • Inspect editability with shape and text statistics
  • Fall back to embedded SVG export when full COM-based conversion is unavailable

How It Works

image / screenshot / rendered PDF
  -> semantic SVG reconstruction
  -> SVG preview and manual correction
  -> PowerPoint conversion
  -> editable PPTX + preview + editability report

Quick Start

  1. Rebuild the source image as semantic SVG.
  2. Preview the SVG render.
  3. Convert the SVG to editable PPTX.
  4. Export a slide preview.
  5. Inspect how editable the output is.
powershell -NoProfile -ExecutionPolicy Bypass -File ".\scripts\render-svg-preview.ps1" `
  -SvgPath ".\out\figure.svg" `
  -OutputPngPath ".\out\figure-preview.png"

powershell -NoProfile -ExecutionPolicy Bypass -File ".\scripts\svg-to-editable-ppt.ps1" `
  -SvgPath ".\out\figure.svg" `
  -OutputPptxPath ".\out\figure-editable.pptx" `
  -WidthPx 1600 `
  -HeightPx 900

powershell -NoProfile -ExecutionPolicy Bypass -File ".\scripts\export-slide-preview.ps1" `
  -PptxPath ".\out\figure-editable.pptx" `
  -OutputPngPath ".\out\figure-editable-preview.png" `
  -Width 1600 `
  -Height 900

powershell -NoProfile -ExecutionPolicy Bypass -File ".\scripts\inspect-pptx-editability.ps1" `
  -PptxPath ".\out\figure-editable.pptx"

Repository Layout

agents/
  openai.yaml
references/
  svg-authoring-rules.md
  workflow-notes.md
scripts/
  render-svg-preview.ps1
  svg-to-editable-ppt.ps1
  export-slide-preview.ps1
  inspect-pptx-editability.ps1
  svg-to-ppt-embedded.js
SKILL.md

Output Quality Goals

A strong result usually has:

  • many PowerPoint shapes
  • nonzero text runs
  • zero embedded pictures in the editable path

When full editability is not possible, the fallback still produces a clean PPTX with the SVG embedded.

Limitations

  • Best for diagrams, flowcharts, figures, and annotated visuals rather than natural photos
  • Full editable conversion depends on PowerPoint COM automation on Windows
  • Very dense icons or decorative artwork may still need approximation or partial raster fallback

Use In Codex

Use the skill when you want to convert a flat diagram image into a semantic SVG and an editable PowerPoint deck:

Use $image2pptx to convert this diagram image into a semantic SVG and editable PPTX.

Roadmap

  • Better automatic diagram decomposition
  • More robust arrow and connector reconstruction
  • More examples and before/after showcases
  • Optional packaging outside Codex skill format

Contributing

Issues and pull requests are welcome. Small reproducible examples are especially helpful for improving editability and visual fidelity.

About

image to ppt

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages