pptx

Write a script using the pptxgenjs library, which is preinstalled, and require it directly without npm install. Set pres.layout before adding slides, use hex colors without the # prefix, and run validate.py on the output to catch chart and slide XML defects PowerPoint would reject.

Unzip the .pptx into a directory, edit the slide XML under ppt/slides/, then rezip from inside the directory. Use add_slide.py to duplicate slides with all package bookkeeping and clean.py to remove orphaned slides, media, and relationships after deletions.

Run markitdown on the .pptx file to get one text block per slide under slide number markers. For a visual overview, run thumbnail.py to generate a labeled grid image of every slide, which helps map content to template layouts.

Common causes include hex colors with a # prefix, negative shadow offsets, outEnd data labels on stacked charts, and combo charts missing valAxes or catAxes entries. Run validate.py on the file; it reports each schema, relationship, and chart failure with its fix.

python-pptx cannot duplicate slides; its only entry point is add_slide(layout). Assigning text_frame.text collapses paragraphs to a single unstyled run, so assign run.text instead, and use add_slide.py from this Skill for slide duplication with full relationship registration.

Yes, .potx templates unpack and pack identically to .pptx; keep the .potx extension on output. Legacy .ppt files must first be converted with the soffice.py wrapper using --convert-to pptx before editing.