Getting Started

Installation

Install the package using npm:

npm install @cognipeer/to-markdown

Or using yarn:

yarn add @cognipeer/to-markdown

Or using pnpm:

pnpm add @cognipeer/to-markdown

The package requires Node.js 18 or newer.

Optional: OCR

OCR is opt-in. If you want to use the default tesseract OCR provider, install its peer dependency as well:

npm install tesseract.js

The vision-model (VLM) OCR providers use the native fetch API and need no extra dependency. See the OCR guide for details.

Basic Usage

Converting from File Path

The simplest way to convert a file is to provide its path:

import { convertToMarkdown } from '@cognipeer/to-markdown';

const markdown = await convertToMarkdown('./path/to/document.pdf');
console.log(markdown);

Converting from Buffer

You can also convert files that are already loaded as buffers:

import { convertToMarkdown } from '@cognipeer/to-markdown';
import { readFileSync } from 'fs';

const buffer = readFileSync('./document.docx');
const markdown = await convertToMarkdown(buffer, {
  fileName: 'document.docx' // Optional but recommended
});
console.log(markdown);

Converting from Base64

Convert base64-encoded data directly:

import { convertToMarkdown } from '@cognipeer/to-markdown';

// With data URL
const base64 = 'data:application/pdf;base64,JVBERi0xLjU...';
const markdown = await convertToMarkdown(base64);

// Or plain base64
const plainBase64 = 'JVBERi0xLjU...';
const markdown2 = await convertToMarkdown(plainBase64, {
  fileName: 'document.pdf' // Helps with format detection
});

Saving Output

Use the saveToMarkdownFile function to save converted markdown to a file:

import { convertToMarkdown, saveToMarkdownFile } from '@cognipeer/to-markdown';

const markdown = await convertToMarkdown('./document.pdf');
const outputPath = await saveToMarkdownFile(
  markdown,
  'converted-document', // filename without extension
  './output' // output directory
);

console.log(`Saved to: ${outputPath}`);

TypeScript Support

The library is written in TypeScript and provides full type definitions:

import {
  convertToMarkdown,
  saveToMarkdownFile,
  type ConverterOptions,
  type ConverterInput
} from '@cognipeer/to-markdown';

const options: ConverterOptions = {
  fileName: 'document.pdf',
  forceExtension: '.pdf'
};

const input: ConverterInput = buffer; // or string
const result = await convertToMarkdown(input, options);

Next Steps

• Learn about converter options
• Explore format-specific conversions
• Read the API overview for the full exported surface
• Browse runnable scripts in the examples/ directory on GitHub