@markitdownjs/core

Core orchestration package for MarkItDownJS. Provides the plugin registry, document pipeline, and renderer — the central package most users install.

Install

npm install @markitdownjs/core

Install converter packages separately as needed:

npm install @markitdownjs/pdf @markitdownjs/docx @markitdownjs/html

Usage

import { MarkItDown } from "@markitdownjs/core";
import { PdfConverter } from "@markitdownjs/pdf";

const parser = new MarkItDown();
parser.registerConverter(new PdfConverter());

const result = await parser.convert({ source: fileBuffer, mimeType: "application/pdf" });
console.log(result.markdown);

Key Exports

| Export | Description | |---|---| | MarkItDown | Main entry point — register converters, run conversions | | MarkItDownOptions | Configuration options for the MarkItDown constructor | | DefaultConverterRegistry | Built-in registry for managing converter plugins | | DocumentPipeline | Low-level pipeline for transforming ConversionInput → AST | | MarkdownRenderer | Renders a DocumentNode AST to Markdown string | | MetadataExtractor | Extracts title, author, dates, and other document metadata | | AssetManager | Handles embedded asset resolution (images, attachments) |

Options

const parser = new MarkItDown({
  includeMetadata: true,
  assetBasePath: "./assets",
});

Result Shape

interface ConversionResult {
  markdown: string;
  ast: DocumentNode;
  metadata: Record<string, unknown>;
}

Part of the MarkItDownJS monorepo.