@fz-labs/pptx-parser-win

Deterministic multimodal PPTX parser that converts PowerPoint files to structured JSON without AI.

Repository: https://github.com/faouzi96/pptx-parser-win

Usage

Install:

npm install pptx-parser-win

Use as a library:

import { parsePptx } from 'pptx-parser-win';

// 1. Write everything (JSON + media files on disk)
const presentationFiles = await parsePptx('presentation.pptx', './output', 'JSON_AND_MEDIA');

// 2. Write JSON metadata only (no image files on disk, images are base64-embedded)
const presentationJsonOnly = await parsePptx('presentation.pptx', './output', 'JSON_EMBEDDED');

// 3. Write absolutely nothing to disk (keep everything in memory, images are base64-embedded)
const presentationInMemory = await parsePptx('presentation.pptx', undefined, 'NONE');

Run as a CLI:

npx pptx-parser presentation.pptx ./output JSON_AND_MEDIA
npx pptx-parser presentation.pptx ./output JSON_EMBEDDED
npx pptx-parser presentation.pptx ./output NONE

CLI arguments

• <pptx-path> — required input file path
• [output-dir] — optional target folder for output files
• [output-mode] — optional output mode (JSON_AND_MEDIA | JSON_EMBEDDED | NONE), defaults to JSON_AND_MEDIA.

Output behavior

• JSON_AND_MEDIA (default): Extracted media files are written to output/media/, slide renders to output/renders/, and presentation.json is written to the output directory. Image elements in JSON contain a path property.
• JSON_EMBEDDED: Only presentation.json is written to the output directory (no image files are created). Image elements in JSON contain an embedded base64 property.
• NONE: No files are written to disk at all. Image elements in the returned Presentation object contain an embedded base64 property.

Package Contents

• main: dist/index.js
• bin: pptx-parser
• exports root entrypoint and CLI from the same file
• compiled files are published from dist/