pdf-figure-vector skill
A skill that turns an academic PDF figure into a self-contained SVG and an editable PPTX, then checks it pixel by pixel.
- 135 ★ on the repository
- Last commit
Minimal access
Why this level
- Scripts start a local preview server and use Chrome for comparison.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/pdf-figure-vectorDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Scripts start a local preview server and use Chrome for comparison.
This is third-party code. Review the repository files before installing.
What it does
pdf-figure-vector converts a chosen PDF figure into vector SVG and a PPTX with editable shapes and text. Elements are rebuilt in the original draw order, fonts are embedded and masked raster inserts get separate handling. The same set can trace ordinary PNG, JPG and WebP files with VTracer. `compare.py` takes a Chrome screenshot and reports the MAE against the PDF render, and `probe.py --bbox` tells a true vector from a raster beforehand.
Who it is for. Graduate students and researchers who need a paper figure as editable graphics for their own slides or preprint.
Good fit when
- A figure from a PDF is needed in a slide where labels and colors can be edited.
- You want to know before starting whether a paper's figure is vector or raster.
- An old PNG block diagram should become an SVG.
- You want to verify the result with a pixel diff against the original.
Not a fit when
- You need an exact copy of a photo or scan: tracing gives an approximation with many paths.
- The figure is copyrighted and reuse is not permitted.
- Chrome is unavailable: the pixel comparison needs it.
Example request
Convert figure 2 from this PDF into vector SVG and PPTX, first check whether it is raster, and show the difference from the original.Limitations
In an SVG produced by image tracing, text becomes outlines and is not editable. If the PDF figure is raster, the skill suggests redrawing it manually instead of passing an image off as vector. Comparison needs Google Chrome and tracing needs the `vtracer` package. The VS Code preview targets Windows and PowerShell, elsewhere the browser opens. The last commit is from October 2026. The author notes that output quality depends on source clarity, so check raster PDFs with `probe.py` first. It is currently a single-maintainer project.
How to disable. Delete the skill folder from `~/.claude/skills/` or `~/.codex/skills/` and stop the local preview server if it is still running.
FAQ
How do I tell a figure is vector?
Run `probe.py` with `--bbox`: it reports whether the figure is native vector or mostly raster. For raster figures the skill proposes manual redrawing.
Are pages numbered from zero?
Yes, the `--page` parameter of `extract_figure.py` takes a page index starting at 0.
What does PNG tracing produce?
An SVG and a PPTX made of native free-form shapes without an embedded full image. The `diagram`, `logo` and `photo` presets balance detail against path count.
Is Chrome required?
Yes, for comparison with the PDF `compare.py` takes a screenshot in Google Chrome on Windows, macOS or Linux.
Where is the result?
In the `--out` folder: SVG and PPTX. Intermediate files are deleted, and `--no-clean` keeps them.
Discussion
No questions or reviews yet.
Related
SkillsRedraws a paper figure as a native Visio .vsdx of editable shapes and text, then exports PNG, SVG, PDF and PPTX from it.
Three skills for paper figures: fig-analyze reads material, fig-draw picks a style and renders, fig-pptx builds an editable PPT.
Skill for publication figures: matplotlib charts for papers in one house style, with vector and raster export from sample scripts.
A skill that draws paper figures as vectors: architecture, mechanism and overview figures exported to SVG, PDF and PPTX
Image to Editable PPT
Image to Editable PPT Skill
A Codex skill that rebuilds slides from images, PDFs and image-based PPTX files into an editable PowerPoint deck
A Codex skill that rebuilds a slide from an image or PDF into PowerPoint where text, tables and arrows stay editable objects.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| pdf-figure-vector (this card) | no | 135 | |
| codex-visio-paper-figure-skill | no | 1,184 | draws scientific schemes in Visio; pdf-figure-vector lifts an existing figure out of a PDF |
| svg-diagram | no | 671 | builds SVG diagrams from a description; here the source comes from a PDF or raster and is verified pixel by pixel |
README in short
The flow is: locate the page and box, check the figure type, extract, compare with the PDF and confirm the PPTX holds shapes. Intermediate files are cleaned up at the end.