OOXML Document Parser (Work in Progress)

A TypeScript library for parsing Office Open XML (.docx, .pptx, .xlsx) documents into structured, page-oriented content.

⚠️ This is a work in progress — APIs and behavior may change. Not yet suitable for production use.

Features

• DOCX: Extracts text split by explicit page breaks (<w:br w:type="page"/>) and embedded images (as Base64 with MIME type).
• PPTX: Extracts slide text, speaker notes, and images (Base64 + MIME type), with slides correctly ordered by filename (slide1.xml, slide2.xml, etc.).
• XLSX: Basic sheet-by-sheet text extraction (tab-separated values).
• Structured output using Content for rich content handling:

  type Content =
    | { type: 'text'; data: string }
    | { type: 'image'; mimeType: string; data: string } // Base64
    | { type: 'note'; data: string }
    | { type: 'comment'; data: string };

Usage

import { parseOoxmlToDocument } from './parser';

const doc = parseOoxmlToDocument('path/to/document.docx');
console.log(doc.pages[0].content);

Limitations

• DOCX: Only splits on explicit page breaks (not layout-based pagination).
• Images: Relies on relationship files; may include unused media if present.
• Comments/Tables/Headers: Not fully supported yet.
• Security: Assumes trusted input; not hardened against malicious OOXML.

Roadmap

• [ ] Add table support (DOCX/PPTX)
• [ ] Extract image alt text and captions
• [ ] Support headers/footers (DOCX)
• [ ] Improve comment extraction
• [ ] Add tests and validation

Use at your own risk. Contributions and feedback welcome!