pptx-to-page

Skill: PowerPoint to Page

Convert PowerPoint presentations to Obsidian Page notes with text extraction, tables, and images.

When to Use

This skill should be invoked when the user:

• Asks to "convert a PowerPoint to a page note"
• Mentions "/pptx-to-page" explicitly
• Wants to import a presentation into the vault
• Needs to extract slides as images for viewing in Obsidian
• Has a .pptx or .ppt file to process

Parameters

• PPTX Path: Path to the PowerPoint file (can be in Downloads, vault, or anywhere)
• Page Title (optional): Custom title for the Page note (defaults to presentation filename)
• Mode (optional): quick (default) or visual
  • quick: Fast docling extraction with text, tables, embedded images
  • visual: Full slide rendering as PNG images (requires LibreOffice)

Process

Step 1: Locate the PowerPoint File

If the user provides a partial path or just a filename:

1. Check ~/Downloads/ first
2. Check vault +Attachments/ folder
3. Ask user for full path if not found

Step 2: Choose Processing Mode

Ask the user (or infer from context):

Quick Mode (default - recommended for most uses):

• Uses docling for fast text/table extraction
• Uses python-pptx for speaker notes and embedded images
• Processing time: ~1 second for 50 slides
• Best for: searchable content, meeting notes, documentation

Visual Mode (when visual fidelity needed):

• Uses LibreOffice to render full slide images
• Processing time: 1-2 minutes for 50 slides
• Best for: design reviews, exact visual reference

Step 3: Quick Mode Processing (Docling)

from pathlib import Path
from docling.document_converter import DocumentConverter
from pptx import Presentation
import os

def process_pptx_quick(pptx_path, output_dir, title):
    """Quick mode: docling + python-pptx extraction"""

    # 1. Docling for text and tables
    converter = DocumentConverter()
    result = converter.convert(pptx_path)
    doc = result.document

    markdown_content = doc.export_to_markdown()
    tables_count = len(doc.tables) if hasattr(doc, 'tables') else 0
    pictures_count = len(doc.pictures) if hasattr(doc, 'pictures') else 0

    # 2. python-pptx for speaker notes and embedded images
    prs = Presentation(pptx_path)
    speaker_notes = []
    embedded_images = []

    for slide_num, slide in enumerate(prs.slides, 1):
        # Extract speaker notes
        if slide.has_notes_slide:
            notes = slide.notes_slide.notes_text_frame.text.strip()
            if notes:
                speaker_notes.append((slide_num, notes))

        # Extract embedded images
        for shape in slide.shapes:
            if hasattr(shape, "image"):
                img = shape.image
                img_filename = f"{title} - Slide {slide_num:02d} - Image {len(embedded_images)+1}.{img.ext}"
                img_path = output_dir / img_filename
                with open(img_path, "wb") as f:
                    f.write(img.blob)
                embedded_images.append((slide_num, img_filename))

    return {
        'markdown': markdown_content,
        'tables_count': tables_count,
        'pictures_count': pictures_count,
        'speaker_notes': speaker_notes,
        'embedded_images': embedded_images,
        'slide_count': len(prs.slides)
    }