Extract a Word document as Markdown, text or blocks
The document on one side, what a pipeline reads on the other.
Loading the example…
The demonstration is not that Markdown comes out — every library in this space can show that. It is the loop: click a word in the extraction and it lights up in the document; click the document and the extraction scrolls to the line it came from.
That join is an address per block, and it is the same address a retrieval system can store and bring back months later. Tables survive as tables, and the list numbers are the document’s own rather than the renderer’s.
import { extractMarkdown } from '@genomdev/genom';
// Headings stay headings, tables stay tables, and the document's own
// list numbers come out as text rather than as CSS.
const markdown = await extractMarkdown(file); // handbook.docx